Haemorrhoids have been diagnosed and treated since the dawn of civilization and yet both their cause and nature, and even their symptomatology, remain hotly debated. The student's path to understanding is further hindered by a confusing terminology, a legacy of folklore, and, most tellingly, anatomical accounts which at best simply ignore pivotal morphological features, and at worst occasionally perpetuate unfounded, though persuasive, myths. Here is a journey to embark upon with both faulty and inadequate maps, and a set of muddling signposts.

The author's task therefore is to guide the reader, for the road is there all right; all that is needed is clearer cartography. He will hope to show that haemorrhoids are not some sort of variocosity, that there is no reason why they should by themselves be particularly itchy or painful, and that the various manifestations of their ‘attacks’ are logically attributable to basic pathological processes; finally he hopes to clarify what exactly the various labels ‘internal’, ‘external’, ‘thrombosed’, ‘strangulated’, and ‘sentinel’, mean.

Piles being a morbid change from the normal, the normal must first be mastered. The anatomy is therefore the key and must be studied in some detail.

The anal canal's interior is usually described as a sphincteric tube, 4 to 5 cm in diameter, lined in its upper two-thirds by mucosa thrown into vertical folds and in the lower one-third by the appendageless squamous epithelium of the sensitive anodermal cuff, the two meeting at the dentate line. This view arises from the routine method of preparing the specimen for dissection. It provides no clue, however, to those intricacies of its infrastructure which explain the almost inevitable, and inherently diverse, vagaries of its average existence.

An anorectum removed from a fresh cadaver and fixed by distension with formalin appears much as described above (Fig. 1) 1134. However, if a specimen is prepared in exactly the same way, but after the veins have been filled (retrogradely through the superior rectal vein), the appearance is quite different (Fig. 2) 1135. The anal lining now bulges as three main pads, more or less subdivided by vertical folds, or ‘columns’ of Morgagni. These are the anal cushions. It is their existence and their curious and unique structure which provide the key to the understanding of piles.

Contemporary anatomical accounts of the anal lining may be inadequate, but many of its special features were noticed more than a century ago. The anal submucosa was likened to cavernous tissue because of its thickness and rich vascularity. Only in recent years, however, was its discontinuous grouping into three main masses observed. These pads, extending above and below the dentate line, are the anal cushions. Morgagni's folds more or less subdivide their mucosal part. Slitting open a specimen prepared as in Fig. 2 1135 shows them en face (Fig. 3) 1136, so belying the traditional anodyne description. At microscopic level these anal cushions exhibit some intriguing features.

The cushions receive a very rich intercommunicating supply from the superior, middle, and inferior rectal (synonymously called haemorrhoidal) arteries. Between five and eight branches of the superior rectal artery pass from the mesorectum through the rectal ampullary wall to descend into the anal submucosa, where they anastomose with the other branches emerging through the muscle wall. Local mucosal excisional procedures inevitably encounter substantial branches from any or all three sources. The profuse arterial supply communicates with the venous system not only through capillaries but also by direct arteriovenous shunts. These provide for a mechanical function (see later) which probably explains the richness of supply.

The veins of the anal lining are distinguished by discrete dilations along their course, particularly below the dentate line in the subanodermal tissues ( Fig. 4 1137,1138 (a, b)). These were once thought to result from disease but are in fact normal, being found in all adults and at birth (Fig. 5) 1139. The veins drain mainly into the superior rectal vein but also through and below the sphincter into the systemic circulation. Dissection preparations suggest that their infrasphincteric communications become increasingly tenuous in adult life (Fig. 4(a)) 1137. If so postdefecatory anal verge engorgement, a problem in some patients, and the oedema and discomfort of some prolapsed piles, may be explained.

The cushions are supported against the shearing, extruding forces of defecation by smooth muscle—the musculus submucosae ani—and by elastic tissue. Discovered by Treitz (1853) and variously described and labelled since, this has for some reason sunk, like the venous saccules, into contemporary oblivion. The musculus submucosae ani descends from the internal sphincter in separate bundles (Fig. 6(a)) 1140 which coalesce beyond the dentate line under the anoderm, to form a dense stroma around the venous saccules there (Fig. 6(b)) 1141. A longitudinal section (Fig. 7) 1142 shows its full extent and demonstrates how the looser submucosal part of the cushions is supported by the tougher more strongly secured anodermal component, and how the muscle's contraction during defecation both flattens the cushions and holds them up against the internal sphincter.

Being composed of a sacculated venous plexus with a rich arterial network the anal cushions provide a spongy variable volume ‘washer’ on which the sphincter can contract, thereby assisting its closure. Their looser textured nature above, supported by the tougher subanodermal part below, is designed to help provide a watertight seal.

The anal cushions can be seen viewed from above in a specimen (Fig. 2) 1135, from below with a proctoscope (Fig. 8) 1143, by transverse histological section ( Figs. 9(a 1144,1145, b)), or by holding the anal lining up to the light (Fig. 10) 1146. The one constant feature is their regular arrangement in the left lateral (3 o'clock), right posterior (7 o'clock), and right anterior (11 o'clock) sectors of the anal circumference: this is, of course, where piles occur. It is logical to conclude therefore that piles are the clinical expression of their internal disruption and downward displacement. It is certainly what they resemble (Fig. 11(a)) 1147, while anal varices, a rare finding, look to be what they are (Fig. 11(b)) 1148.

The anal cushions are disrupted to produce piles by the forces of defecation. For many sufferers defecatory habits and stool consistency are probably to blame. The Valsava effect of excessive straining engorges the cushions, which have lost the support of the external sphincter as it relaxes. The shearing force of hard stools will increase the damage. Other patients, who claim a lifetime of regular easy bowel actions, may have a congenital deficiency in the supporting tissues of the anal cushions. Weakness arising from the influence of progesterone on smooth muscle and elastic tissue probably explains the predisposition to haemorrhoids in pregnancy, although a general increase in pelvic vascularity also contributes. Many women date their haemorrhoids not to actual pregnancy, but to parturition, when the supporting tissues of the anal cushions may be stretched and torn.

Histological examination often shows larger vascular spaces than normal and more prominent connective tissues but no changes not accounted for by the effects of disruption.

The terms ‘internal’, ‘external’ and ‘intero-external’ are often used but serve no great purpose and are confusing, meaning different things to different people. It is better to stick to the basic premise that piles are disrupted cushions and then observe whether the firmly tethered anodermal part is also involved, or whether, as commonly occurs, there is a superimposed skin tag instead.

It is customary to classify haemorrhoids by degree: first degree, only bleeding announces their presence; second degree, spontaneously reducing prolapse at defecation; third degree, prolapse requiring manual replacement; fourth degree, permanent prolapse. However, while a classification is required for the purpose of comparing different treatment techniques scientifically, the degree of any particular patient's piles may vary with time and may be misleadingly represented: indeed one may find gross pile protusion in patients quite unaware of prolapse.

Although the underlying lesion in piles—disruption of the supporting and anchoring tissues of the cushions—means that prolapse is inherent in their nature, patients, are often either unaware of it or unconcerned. Bleeding is much more worrying and is the usual reason for seeing a doctor. Prolapse is, however, the other unequivocal symptom. Pain, itching, and anal dysfunctional effects are less reliable diagnostic criteria.

The capillaries of the lamina propria are only protected by a single layer of epithelial cells, and little trauma is required to breach them. Since it is the more lax-textured, upper part of the anal cushion which mainly prolapses, dragging the mucosa to the outside, trauma due to wiping or contact with clothes often occurs. Repeated trauma produces a chronic inflammatory response, making the damaged mucosa a brighter red and granular (Fig. 12) 1149, and so more friable and likely to bleed.

A great deal of unnecessary investigation, which is costly, inconvenient, uncomfortable, and occasionally even hazardous for the patient, can be avoided by time spent unravelling exactly what is meant by bleeding per rectum. Patient and courteous attention to detail in taking a history is always amply repaid, but never more so than in anal bleeding: haemorrhoids are very common, and yet bleeding may also be indicative of a more serious condition. First and second degree piles, which remain intra-anal except at defecation, bleed with the bowel movement. Being capillary blood it is bright red. If enquiry reveals that it occasionally drips, an anal origin is certain, because the anus remains closed by tonic contraction of the sphincter except at the moment of defecation. Blood that drips into the pan, after passage of the stool, must originate from extruded anal mucosa, or from a fissure in the anoderm. The only other, and extremely uncommon, possibility is a rectal polyp on a long enough stalk. Blood smeared on the stool in the pan is ominous and unlikely to be coming from piles, since freshly shed blood ought to disperse straight away into the water. The fact that it remains on the stool suggests either that it has congealed there, or is mixed with mucus, indicating a higher lesion.

Passage of clotted blood also demands exculpation of a colorectal source, and a careful history may provide a useful clue. Piles may still be the explanation if questioning reveals that the clots were only seen on the paper and looked fresh; such clotting can have occurred in freshly shed blood lying at the anal verge. It is very rare for a large pile to bleed back into the rectum and proclaim itself by passage of older clots at stool. Third and fourth degree piles may bleed into clothing, sometimes dramatically. Occasionally a patient's claim to passing dark blood is found to derive from seeing it on clothing after it has dried and deepened in colour.

Patients may be quite unaware of protruding anal cushions, even when they are fourth degree, and surprisingly few report for treatment for this symptom alone. Descriptions of a fleshy lump or prolapse usually have to be elicited by questioning. For some, however, life is plagued by a pile which prolapses on exertion, making them uncomfortable, exuding mucus, and generally restricting their work or leisure. Others have to lie down after defecation to reduce the protusion.

Pain is a contentious issue in pile symptomatology. Although claimed to be a prominent and attributable problem, there seems to be no good reason why a disrupted anal cushion should actually be painful. When trapped outside the closed anus, distortion combined with oedema and congestion from lymphatic and venous impairment may well cause discomfort. In many causes pain on defecation is due to an easily overlooked fissure. Some patients nevertheless do experience relief from pain after simple treatment of their undoubted piles.

Episodes of painful irreducible swelling which last a week or so can be most unpleasant. Often given the inaccurate epithet ‘strangulated’ piles, they are usually due to greater or lesser degrees of infarction resulting from obstruction of venous drainage by thrombosis and consecutive clotting in the sacculated venous plexus. ‘Infarction’ is used here in its proper sense, denoting an intravascular and interstitial ‘stuffing with blood’, and not in its common contemporary misusage implying necrosis. Although necrosis would supervene if circulatory impairment by venous blockage were sufficient, the usual outcome is spontaneous resolution as the clot shrinks and lyses and venous circulation is restored. This condition is discussed more fully below.

When the patient's main concern is itching, piles are seldom, if ever, to blame: a local skin condition is usually responsible. Although treatment of coexisting piles may help relieve itching it is unwise to encourage a patient greatly bothered by pruritus to believe that relief will be produced from curing piles.

Defecatory derangement can result from excitement of a disrupted anal cushion, causing a sensation of incomplete evacuation, particularly when the cushion is engorged by fruitless straining, which further congests the cushions and worsens the problem. Unsatisfied defecation is also a feature of other anorectal conditions, including rectal tumour.

Blood and serum from the exposed inflamed mucosal part of a pile dries dark on underclothing and may be thought faecal. Only very rarely, however, do third and fourth degree haemorrhoids allow minor conduction of rectal contents to the surface. Mucus may also exude from the exteriorized mucosa of piles and can be the presenting symptom.

When a meticulous history suggests piles and the findings agree examination can be confined to the anorectum. The only equipment required is proctoscope, sigmoidoscope, light source, and biopsy forceps.

There are several dynamic influences on a pile's presentation—the vigorous arterial supply, the presence and possibly changing diameter of the arteriovenous shunts, the variability of cushion bulk due to capacity of the venous saccules, and the effects of cushion displacement and anal sphincter contraction on venous and lymphatic drainage. As a result, not only does the appearance change from time to time in the same patient, but the same symptom may have different causes in different patients. For instance, whereas most patients complaining of prolapse have a simple displaced pile (Figs. 11(a), 12) 1147,1149 a ‘lump’ felt by others may be due to engorgement of the subanodermal veins due to impairment of drainage (Fig. 13) 1150 or transient post-defecatory anodermal oedema (Fig. 14) 1151.

Piles that are transiently displaced suffer little trauma, but when the mucosal part is frequently exposed it becomes inflamed (Fig. 12) 1149. Thrombosis and clotting in the vein sacs (when blood flow may be turbulent) also influence the appearance of the pile, but as an indication there will be associated discomfort or, depending on the extent of clotting and consequent infarction, frank pain. Solidification of a small part of the venous plexus causes an uncomfortable attack of swelling of the pile, with oedema but little infarction. Greater degrees of obliteration of venous drainage embarrass the circulation accordingly (Fig. 15) 1152. The fully infarcted pile (Fig. 16) 1153, despite its appearance, usually settles uneventfully. Many patients who seek medical attention because of such an attack of saccular clotting, and who graphically describe the severity of the condition, have recovered by the time of specialist consultation. The term ‘strangulated piles’ may be misapplied to this condition, causing inappropriate and inevitably unsuccessful efforts at supposedly therapeutic replacement.

A disordered cushion may, therefore present in one of several ways as a lump at the anal verge. In most cases, however, external inspection provides no clue to their presence, and nothing abnormal is found on anal digitation, since uncomplicated piles are impalpable. A nodular induration is felt if clotting has occurred and this may occasionally become firm due to fibrosis. In most patients, the diagnosis is suggested by the history and confirmed by proctoscopy. Interpretation of the proctoscopic appearance is not straightforward. Since anal cushions are normal structures (Fig. 8) 1143, their distinction from piles (Fig. 17) 1154 is only one of degree. Bright red granularity of the mucosal part of a cushion is certain evidence of its disruption, and the extent to which the cushions bulge into the instrument's end on straining and follow it out on withdrawal, provide a valuable guide.

Sigmoidoscopic exclusion of rectal disease is an essential part of the establishment of the diagnosis. Because piles are common finding them does not rule out another condition higher in the rectum causing the symptoms. There is, however, no evidence for the claim still occasionally made that haemorrhoids can result from rectal carcinoma or pelvic masses (an erroneous idea originating from the belief that piles resulted from venous obstruction).

Many patients mistake anal tags for piles, and indeed the disrupted anodermal part of a cushion may have a similar appearance. Anal tags are cutaneous protruberances at the junction of the anoderm and perianal skin. They are of uncertain origin, but possibly result from local derangement of lymphatic drainage, as their occasional disarming partial reformation soon after excision suggests. They can be solitary and discrete, or form a circumferential irregular fringe (Fig. 18) 1155.

These are club-like protruberances from the dentate line and seem to be hypertrophied anal papillae, again possibly due to lymphatic obstruction (Fig. 19) 1156.

A patient described as having ‘painful itchy piles’ may well be suffering from an anal fissure, particularly if this is associated with a sentinel tag masquerading as a pile. The deep burning pain of a fissure on and after defecation and the associated itching are quite unlike the discomfort which might accompany a pile.

Eczematous, psoriatic, and fungal dermatitis causes anal discomfort which often promotes referral as a case of haemorrhoids. Hyperkeratosis (seen as pale slightly soggy or glazed skin), punctate excoriations, and hairline radiating skin cracks will suggest the correct diagnosis.

Perianal haematoma is one of the many local misnomers applied to anal disorders. It appears as a dark, tender, berry-like lump at the anal verge which is in fact a clotted subanodermal venous saccule (Fig. 21) 1158. Its other name, also inaccurate, is ‘thrombosed external pile’, a term which, to add to the confusion, is also applied to infarcted piles.

Early rectal prolapse may be confused with piles. If the patient is inhibited from straining sufficiently to produce the lesion, the prolapse may look like an anterior pile. Full prolapse, however, has an unmistakable appearance. It is rare in men.

Rectal tumours can easily be missed by insufficient attention to the history and a careless digital examination, for it is not so much the length of the finger which matters, as the amount of thought behind it. Because of the rectum's curvature even upper-third tumours may be palpable. Even when nothing is felt or seen, if the patient's symptoms do not accord with the findings, further investigation is required. Ominous symptoms are old blood, particularly if slimy or clotted, unsatisfied defecation, deep discomfort, and ‘wet’ flatus.

Piles should be treated on their merits, tailoring the procedure to the patient's predicament and complaint. Only rarely are piles a threat to health in causing anaemia, so treatment should be suggested, not urged, and directed by symptoms rather than the attendant's obsession with neatness. Surgery is the replacement of one lesion by another; the aim is to ensure that the second is preferable to the first.

Management is either conservative or interventional, the latter being divided into those techniques aimed at reducing tissue volume and promoting adhesion of the remainder, and those claiming to work by other means.

Only by returning to the principle that piles are disrupted anal cushions can the beginner make any sense of the bewildering array of treatments available. At the dawn of this era of diversity, the Lancet aptly mocked the dilemma produced by the advent of yet another proclaimed method by heading the editorial: ‘To tie, to stab, to stretch; perchance to freeze’.

As the fluctuating severity and intermittent nature of symptoms suggest, piles in their early stages are not an all-or-nothing complaint. For many patients advice to increase dietary fibre to bulk and soften stools, removal of literature from the lavatory, avoidance of excessive straining, and prompt manual replacement, may be sufficient to prevent marginally disrupted cushions from being much of a nuisance. Bulk laxatives may be helpful, as may simple suppositories, which certainly lubricate the adjacent surfaces so that the piles replace more readily and may in addition have some unknown emollient effect on the piles themselves, as their protagonists believe.

This works by strangling a ‘polyp’ of the insensitive mucosal part of a pile with a small elastic O-ring. It is mounted on a ligating device and snapped on to the pile through an anoscope by a releasing trigger, the grasped tissue having been first pulled into place. The strangled tissue withers and falls away in a few days, leaving a small ulcer which usually heals within a month. No anaesthesia is required.

Both of these cause tissue destruction by heat. The instruments are applied through an anoscope to coagulate a predictable volume of adjoining tissue. The mucosal part is treated; no anaesthesia is required.

An irritant chemical solution, usually 3 ml of 5 per cent phenol in arachis oil, is injected into the submucosa at the base of each pile. When the varicose vein theory of piles prevailed it was thought to act by inducing fibrosis to constrict the superior rectal venous drainage, so, it was thought, deterring transmission of the supposed high pressures from the portal system. In fact it probably causes some shrinkage of tissue by necrosis, and adhesion as a result of the ensuing inflammatory reaction.

A liquid nitrogen probe is placed against the pile for 3 min, and causes cold necrosis. Local anaesthesia at least is said to be needed.

Here, the offending bulk is excised. ‘Open’ and ‘closed’ methods are used. The ligation/excision technique of Milligan and Morgan, which is safe and well tried, leaves an open wound at the site of the pile, the ligated pedicle of which, containing branches of the superior rectal artery, lies alongside. When the pedicle has separated the wound heals by secondary intention.

Primarily closed wounds however can be safely achieved both by the submucosal dissection haemorrhoidectomy, whereby excess tissue is excised from under raised mucosal and anodermal flaps which are then stitched back, or simply by diathermy excision of the pile within a longitudinally disposed ellipse, achieving haemostasis by diathermy coagulation and then wound closure with absorbable stitches.

This was widely practised as a treatment for piles in the last century. The nineteenth century French surgeon, Verneuil, again reconciling its effect with the varicose vein theory, thought it improved venous drainage by stretching the rectal muscular button holes which convey the anal tributaries of the superior rectal vein. When reintroduced some years ago, it transiently displaced the surgical standby of the time, haemorrhoidectomy, but was later shown to have limited application. It probably helps patients who have discomfort and difficulty in defecation by easing the effort of evacuation, so reducing the congestion of the cushions from excessive straining.

This has been advocated. Absorbable sutures are placed above the dentate line to attach the cushion back to the internal sphincter. Obliteration of its blood supply also reduces bulk.

Pain is inevitable after haemorrhoidectomy, whatever the technique. Band ligation also causes pain, but this is unpredictable and many patients suffer little discomfort. Clotting in the adjacent sacculated venous plexus is occasionally precipitated by banding, causing severe, prolonged pain. Combining banding with an anal stretch seems to increase this risk and is not recommended. Because of pain and the possibility of fainting it is sensible to warn the patient in advance against driving home unaccompanied after such outpatient procedures.

All methods except injection, stretching and perhaps haemorrhoidectomy by primary closure techniques leave a moist open wound to heal by secondary intention and thereby carry the irreducible risk of secondary haemorrhage. Although rare it is most worrying for the patient. Anxiety is reduced by explaining the possibility beforehand and the methods by which it can readily be managed—traction on a balloon catheter inflated in the rectum, say, or suturing the bleeding point. As a wise precaution, procedures carrying a risk of secondary haemorrhage should not be undertaken within 3 weeks of a trip abroad or to remote parts.

Sepsis seldom complicates either haemorrhoidectomy or ambulatory procedures. When it occurs it is managed with antibiotics and drainage where appropriate. Haemorrhoidectomy, even by banding, can be followed by infection in the immunocompromised patient.

Over-vigorous stretching, particularly in the elderly, can have this disastrous consequence. The procedure should always be performed in a controlled way using the finger tips, gauging the force used to the strength of contracture of the tissues, rather than according to an arbitrary numerical formula. The stretch should be distributed equally around the circumference by changing direction.

Given the anal cushion's function, removal of excess tissue during haemorrhoidectomy can be expected to impair continence, particularly since most of the tissue excised is anodermal. When the procedure is necessary, therefore, the amount of tissue excised should be kept to the minimum compatible with relief of symptoms.

Haemorrhoidectomy is notorious for causing transient difficulty in voiding in men. Banding may also induce curious bladder symptoms.

A generation ago the choice was simple: piles classed as first or second degree were injected, and third and fourth degree cases underwent haemorrhoidectomy. Little was done properly to test the efficacy of either. With the introduction of ambulatory tissue-reducing techniques and the advent of the now ubiquitous clinical trial, the scene has entirely changed. Haemorrhoids make an excellent subject for trials. The supply is inexhaustible and the effects of treatment can be assessed both subjectively and objectively. There are also many different treatments to evaluate.

Conservative treatment has been compared with band ligation; bulk purgatives or high fibre diet with sclerotherapy, stretching, sphincterotomy, band ligation, and freezing. Banding has also been compared with sclerotherapy, haemorrhoidectomy, stretching, and infrared photocoagulation. Photocoagulation has been compared with bipolar diathermy, and one type of haemorrhoidectomy has been judged against another. Excellent and praiseworthy though such trials are they suggest a ‘rivalry’, when in fact no one treatment is best for all patients. Their greater merit is in showing prospectively and with careful control and monitoring, that benefit can be derived from each measure, so allowing us an informed choice of the possible treatments.

Haemorrhoidectomy still has a place when gross disruption of the sensitive anodermal part of the cushion is the main problem. A trial has shown that simple ligation/excision is in fact no more painful than submucosal dissection. When the shape and size of the pile is appropriate, closed methods of haemorrhoidectomy have much to recommend them. The obvious drawback of haemorrhoidectomy is its requirement for admission, anaesthesia, and a longer recovery time. Cryodestruction is a messy, prolonged, and anatomically less accurate way of achieving the same end as haemorrhoidectomy. Infrared coagulation has the same beneficial effect as banding. Since it causes less pain and requires fewer days off work, it has greater patient acceptability. Band ligation, however, achieves its objective in significantly fewer treatment sessions and the instrument is both a great deal cheaper and more robust. The discomfort of banding can be rendered more tolerable, particularly in the light of more immediate benefit, if the patient is adequately warned beforehand and provided with analgesic tablets, and something to keep the motions soft. Forcible anal dilation may have a limited adjunctive place in the treatment of piles but would no longer be advanced as an alternative management. Sclerotherapy can be effective, but the improvement is not so reliably maintained as with other methods. Its undoubted advantage is cost and convenience.

Although some clinicians advocate emergency anal stretching and others haemorrhoidectomy, in the author's experience the majority of cases settle spontaneously. Previous symptoms improve, presumably due to fibrosis. Management therefore is simply bed rest as required, stool softeners, and analgesia. Local anaesthetic preparations have variable effects but may be helpful. Surgery in areas of such compromised tissue may be unwise and excessive, and recovery is no quicker. In the rare event of infarction progressing to necrosis, however, and in the occasional patient suffering severe and prolonged pain, debridement haemorrhoidectomy (cutting away the dead tissue) speeds recovery and certainly relieves the pain dramatically.

Once one understands the detailed anatomy of haemorrhoids and their pathological possibilities, their management is straightforward, and the treatment effective, acceptable, and reasonably trouble free. The important part is diagnosis. The patient's preconception that piles are responsible for their symptoms must not cloud the clinician's mind. Uppermost must be the question whether reducing the bulk of the anal cushions and mooring them more firmly will logically address the presenting problem. If not, then in a 19th century surgeon's words, ‘in such cases prudence equally forbids the rash interposition of unavailing art, and the useless indulgence of delusive hope’.

Ambrose NS, Hares MM, Alexander-Williams J, Keighley MRB. Prospective randomised comparison of photocoagulation and rubber band ligation in treatment of haemorrhoids. Br Med J 1983; 286: 1389–91.

Dehn TCB, Kettlewell MGW. Haemorrhoids and defaecatory habits. Lancet 1989; i: 54–5.

Dennison AR, Whiston RJ, Rooney S, Morris DL. The management of haemorrhoids. Am J Gastroenterol 1989; 84: 475–81.

Dennison, A, Whiston RJ, Rooney S, Chadderton RD, Wherry DC, Morris DL. A randomised comparison of infrared coagulation with bipolar diathermy for the outpatient treatment of haemorrhoids. Dis Colon Rect 1990; 33: 32–4.

Editorial. To tie, to stab, to stretch, perchance to freeze. Lancet 1975; ii: 645.

Gibbons CP, Trowbridge EA, Bannister JJ, Read NW. Role of anal cushions in maintaining continence. Lancet 1986; i: 886–8.

Greca F, Hares MM, Nevah E, Williams JA, Keighley MRB. Randomised trial to compare rubber band ligation with phenol injections for treatment of haemorrhoids. Br J Surg 1981; 68: 250–2.

Haas PA, Fox TA, Haas GP. The pathogenesis of haemorrhoids. Dis Colon Rect 1984; 27: 442–50.

Hosking SW, Smart HL, Johnson AG, Triger DR. Anorectal varices, haemorrhoids, and portal hypertension. Lancet 1989; i: 349–52.

Jensen SL, Harling H, Arseth-Hansen P, Tange G. The natural history of symptomatic haemorrhoids. Int J Colorect Dis 1989; 4: 41–4.

Marshman D, Huber PJ, Timmerman W, Simonton CT, Odom FC, Kaplan ER. Haemorrhoid ligation. Dis Colon Rect 1989; 32: 369–71.

Murie JA, Mackenzie I, Sim AJW. Comparison of rubber band ligation and haemorrhoidectomy for second and third degree haemorrhoids. Br J Surg 1980; 67: 786–8.

Murie JA, Sim AJW, Mackenzie I. The importance of pain, pruritus and soiling as symptoms of haemorrhoids and their response to rubber band ligation. Br J Surg 1981; 68: 247–9.

Roe AM, Bartolo DCC, Locke Edmunds J, Mortensen NJ McC. Submucosal versus ligation excision haemorrhoidectomy. Br J Surg 1987; 74: 948–51.

Templeton JL, Spence RA, Kennedy TL, Parks TG, Mackenzie G, Hanna WA. Comparison of infrared coagulation and rubber band ligation for first and second degree haemorrhoids: a randomised prospective clinical trial. Br Med J 1983; 286: 1387–9.

Thomson WHF. The nature of haemorrhoids. Br J Surg 1975; 62: 542–52.

Thomson WHF. Non-surgical treatment of haemorrhoids. Br J Hosp Med 1980; 24: 298–301.

Thomson WHF. The one-man bander: a new instrument for elastic ligation of piles. Lancet 1980; ii: 1006–7.

Thomson WHF. The real nature of ‘perianal haematoma’. Lancet 1982; ii: 467–68.

Senapiti A. A randomised controlled trial of injection sclerotherapy with bulk laxatives. Int J Colorect Dis 1988; 2: 124–6.