ACTINOMYCOSIS

A case of sigmoid actinomycosis is described and the cases of abdominal actinomycosis presenting at the Manchester Royal Infirmary over the past 20 years are reviewed. The whole problem of abdominal actinomycosis is discussed, with special reference to the sites of infection and mode of treatment. The relevant literature is reviewed.

A case of sigmoid actinomycosis is described and the cases of abdominal actinomycosis presenting at the Manchester Royal Infirmary over the past 20 years are reviewed. T h e whole problem of abdominal actinomycosis is discussed, with special reference to the sites of infection and mode of treatment. T h e relevant literature is reviewed.

THE first clinical description of actinomycosis has been attributed to Von Langenbeck (1845). In 1877
Bollinger demonstrated that a fungus was the infective agent in producing the hard masses that occurred in the jaw bones of cattle. T h e following year Harz (1878) suggested the name Actinomyces bovis for this fungus. At the same time Israel (1878) was studying material obtained from a patient with suppuration in the neck.
He noted granules which contained the same mycelia as those described by Bollinger (1877) in cattle. Israel, who remained active in the study of actinomycosis, considered that it is an endogenous anaerobic infection and despite suggestions that it is an exogenous aerobic infection by such workers as Bostroem (1890) this theory has been upheld ever since. About 20 per cent of cases of actinomycosis have the primary site of infection in the abdomen (Cope, 1938), the majority originating in the region of the caecum. Other less. common sites are the stomach, duodenum, gemto-urmary tract, and colon. T h e sigmoid colon is a very rare site of infection, and this is why we decided to report the present case.

REVIEW OF C A S E S
During the past 20 years 7 cases of abdominal actinomycosis have been treated at Manchester Royal Infirmary. T h e total number of cases of actinomycosis treated at the same hospital during this period was 13. This represents an unusually high proportion with the primary site of infection in the abdomen, which is probably explained by the fact that the cervical cases would most probably have been referred to a dental hospital. Of the 7 cases, 5 originated in the appendix, I in the sigmoid colon, and I in either the gall-bladder or the duodenum. I n all these cases the diagnosis was confirmed by growth of the actinomyces on culture or by the discovery of the typical colonies in histological sections. Suspected cases not confirmed by either histology or bacteriology have not been included.

CASE REPORTS
Case I.-A 29-year-old housewife was admitted to Manchester Royal Infirmary in September, 1970, complaining of pain in her left iliac fossa which had persisted for 4 months. She had lost about 3 kg. in weight but there was no change in bowel habit. Her only other complaint was of mild frequency of micturition. On examination she was very pale with a temperature of 38" C . Palpation of her abdomen revealed a firm tender mass, 10 cm. in diameter, in her left iliac fossa. The mass was fixed to the left side of her pelvis and rectal examination revealed a tender swelling in the pouch of Douglas on the left side. The haemoglobin was 10.3 g. per cent and the white-cell count was 12,000 c.mm. with a normal differential. A midstream specimen of urine showed no pus cells.
A laparotomy was performed on 14 Sept. through a left lower paramedian incision. A fixed mass was found in the left iliac fossa adherent to the side-wall of the pelvis and involving the sigmoid colon. The left Fallopian tube and ovary were not involved, but the transverse colon and omentum were tethered to the upper margin of the mass. The mass itself was composed of dense fibrous tissue and contained multiple abscesses. The nature of the pathology was not clear from the macroscopic appearance. The abscesses were drained and specimens of the pus were sent for culture. Portions of the walls of the abscesses were sent for histological examination by frozen and parsections. The frozen sections were reported as showing non-specific acute inflammation. The wound was closed :nd the area drained through a separate stab incision. No sulphur granules' were seen at this time. Bacteriological studies showed Gram-positive cocci which failed to grow on culture. The biopsies revealed acute abscesses with surrounding granulation tissue and mature fibrous tissue. No specific granulomata were seen.
Postoperatively the patient was given 2 pints of blood and started on penicillin, I MU. 6-hourly, and streptomycin, 0.5 g. twice daily intramuscularly. As bacterial cultures were sterile antibiotics were discontinued after 5 days, by which time her temperature was normal. The she had developed a swelling in her right iliac fossa which wound healed well and she was allowed home 10 days had started to discharge after a week. On examination later on no specific treatment. She was seen for follow-up she was found to have a mass, 8 cm. in diameter, in her on 6 Nov., when she still complained of a dull ache in her right groin which was discharging yellow pus. No left iliac fossa. Her haemoglobin and white-cell count were normal at this stage and a barium-enema examination, although showing no specific abnormality, revealed some rigidity and atypical irregularity in the sigmoid colon.
The radiologist suggested that she might have Crohn's disease. She attended again on 11 Dec., when she complained of a recurrence of the swelling in her left iliac fossa with a discharge from the site of her previous drain. She was started on ampicillin, 250 mg. four times a day orally, and was seen a week later when her condition was unchanged. One month later the mass was larger and she was admitted to hospital for the second time. She was again pyrexial with a temperature of 37.8" C., but her white-cell count was only 10,300 c.mm. Her haemoglobin had dropped to 12.1 g. per cent. Her erythrocyte sedimentation rate was 95 mm. in the first hour. A colonoscopy was carried out and no abnormality was seen to the limit of the examination. The instrument could not, however, be passed beyond the distal sigmoid colon owing to the rigidity of this part of the bowel. A second operation was performed on 28 Jan., 1971, when the sinus at the previous drain site was excised and an abscess involving the full thickness of the abdominal wall was opened. Yellow granules were noted in the pus and a clinical diagnosis of actinomycosis was made. The abscess was drained and after multiple biopsies had been taken the wound was closed around a corrugated drain (Fig. I). The clinical diagnosis was confirmed both on culture and histological examination (Figs. 2, 3). The organism was found to be sensitive to penicillin, tetracycline, and erythromycin. She was therefore started on long-term penicillin therapy, I MU. 6-hourly intramuscularly for the first week, I MU. 12-hourlyintramu~cularly for I week, followed by I MU. daily intramuscularly as an out-patient for I month. This was supplemented by oral penicillin 'V', 250 mg. four times a day for 2 months. Her erythrocyte sedimentation rate, which was 111 mm. per hour at the time of the second operation, returned to normal limits after 2 weeks of treatment and after 2 months of treatment all traces of the mass had disappeared and the sinus had not recurred. Her liver-function tests at the height of the infection had shown a reversal of the albumin/globulin ratio (albumin, 2.8 g. per cent; globulin, 5.1 g. per cent), suggesting involvement of the liver by the fungal infection. At the end of the course of treatment her liver-function tests had returned to normal.
Case 2.-A female, aged 21 years, presented with a 2-month history of pain in her right groin preceded by 3 months' general malaise. Three weeks before admission granules were found in the pus which on first culture failed to reveal Actinomyces. Her haemoglobin was 7 g. per cent and her white-cell count was 15,ooo c.mm. with 89 per cent of polymorphs. At operation the abscess was drained and culture of the pus revealed actinomycosis. She was given a 2-month course of soluble penicillin, 2 MU. daily, following which the sinus healed. Case 3.-A female, aged 27 years, gave a long history of recurrent attacks of small-bowel obstruction for which she had had three operations prior to her presentation in May, 1951. Examination of her abdomen at this time revealed abdominal distension with a palpable mass in her right iliac fossa. At laparotomy there was a mass in the caecum. Biopsies were taken and an ileotransverse anastomosis was performed. The biopsies revealed no specific abnormality and a provisional diagnosis of either Crohn's disease, or carcinoma, or tuberculosis was made. In November of the same year she returned with a discharging sinus in her right iliac fossa and at this time 'sulphur granules' were seen in the pus. A diagnosis of actinomycosis was made. She was therefore started on soluble penicillin, I MU. daily intramuscularly. Unfortunately, she died shortly afterwards.
Case 4.-A female, aged 30 years, presented with a &week history of general malaise and pain in the right iliac fossa. On examination there was a palpable mass in her right iliac fossa. At laparotomy the appendix was found to enter a mass which contained a moderate amount of pus. The abscess was simply drained and she was started on I MU. of penicillin twice daily intramuscularly which was continued for 2 months. Culture of the pus confirmed the diagnosis of actinomycosis. The sinus cleared up permanently after 3 months.
Case s.-A male, aged 41 years, presented in 1953 with history of pain in his right hypochondrium and dyspnoea. His temperature was 38.5"C. and a diagnosis of acute cholecystitis was made. Subsequently, however, the gallbladder was found to function normally on radiography. Whilst in hospital he developed a fluctuant swelling in his right flank which was thought to be a perinephric abscess. This was drained and culture of the pus grew a nonhaemolytic streptococcus. He was put on streptomycin intramuscularly. In 1957 he developed a recurrence of the pain in his right hypochondrium and was put on the waiting list for cholecystectomy. Before this could be done he was admitted as an emergency with an area of cellulitis over his right costal margin. He was put on penicillin therapy and the area was aspirated on three occasions with negative culture. Following this, the area was explored and the abscess drained. Culture on this occasion grew only a non-haemolytic streptococcus. In 1959 the abscess recurred. It was explored more thoroughly and was found to track subdiaphragmatically. Culture of the pus on this occasion revealed Actinomyces. He was put on repeated courses of penicillin intramuscularly. After three further flare-ups of the infection the condition subsided.
Case 6.-A male, aged 17 years, had an appendix abscess drained in November, 1953, at which time it was impossible to remove the appendix. In January, 1954, an operation was performed to remove the appendix which was adherent to the bladder. The appendix was successfully removed but the wound discharged yellow pus postoperatively. The discharge persisted until May, 1954, when another abscess developed in his left lumbar region. This was drained and revealed thick yellow pus which grew Escherichia coli, non-haemolytic streptococcus, and Staphylococcus aureus. Without antibiotic therapy the discharge from his appendix wound and the lumbar abscess subsided. In July, 1954, he developed an abscess in his sacral region. This abscess was aspirated and revealed Actinomyces. He was started on penicillin, I MU. intramuscularly twice daily. After I month of treatment the sinus finally healed.
Case 7.-A male, aged 13 years, gave a 2-week history of diarrhoea following which he had developed an abscess in his right groin. A diagnosis of inguinal lymphadenitis was made and he was started on penicillin 600,000 U. daily. The abscess failed to subside and I month later was drained, revealing Actinomyces. Streptomycin was added to the penicillin therapy and the abscess cleared up. Three years later he was admitted with acute appendicitis. At operation the appendix was found to be adherent to the scar of the previous drainage wound in the groin. Following the removal of the appendix he made a good recovery.

DISCUSSION
Pathogenesis of Abdominal Actinomycosis.- The organism responsible for abdominal actinomycosis is an anaerobic, Gram-positive, branching filamentous fungus which is related in many respects to bacteria. Actinomyces israeli is a common saprophyte in the mouth and gastro-intestinal tract, the endogenous nature of the infection being virtually beyond doubt. Slack (I942), who examined IOO pairs of tonsils from routine tonsillectomies, found colonies of anaerobic actinomycetes in 14 per cent of cases. T h e organism enters the abdominal cavity from the alimentary tract in the majority of cases and infection from the genital tract, diaphragm, or blood-stream is extremely rare (Cope, 1938). T h e actinomycetes cannot penetrate healthy mucosa and damage to the mucosa by trauma, disease, or surgery is an essential prerequisite to infection. Having penetrated the bowel, the organism proliferates, leading to the production of a mass of dense, almost woody, fibrous tissue. Areas of this mass usually undergo softening with the production of abscesses, which may rupture through the abdominal wall producing sinuses, penetrate into adjacent viscera causing fistulae, or rupture through the diaphragm leading to pulmonary actinomycosis. Lymphatic spread in actinomycosis does not occur, presumably because of the large size of the filaments; lymphadenitis is due to the secondary infection which is very common. Blood spread, however, is quite common and hepatic involvement via the portal vein leads to the typical 'honeycomb ' liver.
Actinomycotic pus consists of a dense infiltration of mainly polymorphs, in the centre of which are colonies of actinomycetes forming the ' sulphur granule'.
This sulphur granule or mycelium tends to be surrounded by club-shaped excrescences which radiate outwards, giving rise to the alternative name of 'ray fungus'. These clubs are not formed in artificial culture and are produced by deposition of host lipid on to the mycelium (Walter and Israel, Abdominal Localization.-Apart from the neck, the abdomen is the commonest site for actinomycosis, accounting for one-fifth of the cases (Cope, 1938). T h e ileocaecal region is the usual site as a result of a perforated appendix, but no site is immune.
Stomach and Duodenum-Perigastric actinomycosis occurs only when the gastric mucosa is diseased or after gastric surgery. T h e actinomycetes cannot survive in a healthy stomach (Cope, 1938). Nathan (1929) reviewed gastric actinomycosis and described a fatal case which was found to have a large callous ulcer penetrating into the liver and pancreas. H e attributed the rarity of the infection in this site not to gastric acidity but to vigorous motility preventing the fungus from settling. Cope (1938) believed that perforation of the mucosa was essential. Perigastric actinomycosis occurs in two circumstances: either associated with a perforated ulcer or following gastric surgery. A common feature in all the cases described is a latent period before the onset. Till (1946) described a case presenting 4 months after a perforated gastric ulcer. More recently Pheils, Reid, and Ross (1964) and Ashton and Slaney (1963) have described cases following gastric surgery. T h e infection in Ashton and Slaney's case complicated a leaking duodenal stump. Again a long latent period was noted. A rare complication of gastric actinomycosis was reported by Wilson (1961)~ who described pyloric stenosis following infection in the antrum. T h e stenosis responded dramatically to parenteral penicillin and sulphadiazine.
Gall-bladder and Pancreas.-Invasion of the pancreas and biliary tree by the fungus is extremely rare. T h e pancreas most probably becomes infected by direct spread from the stomach or duodenum owing to penetration by a peptic ulcer. Fontan, Serrano-Rios, Navarro, Oliva, and Ramirez (1969) described a fatal case in a 62-year-old man who had multiple actinomycotic abscesses in the pancreas and liver. T h e present Case 5 probably represents a case of actinomycotic cholecystitis and we have found only 1963).
z other established cases in the literature (Mayo-Robson, 1905;Rewerts, 1964). T h e mode of infection of the gall-bladder by the fungus is less well understood. Mayo-Robson (1905) felt that infection did not occur via the skin, blood-stream, or lymphatics or by direct spread from the gut as no other lesion was present in his case. He believed that the infection reached the gall-bladder by retrograde spread up the bile-duct from the duodenum. Cope (1938) pointed out that once the organism has penetrated the bowel the mucosa heals and may show no sign of the initial infection, so that this is the most likely route of infection in actinomycotic cholecystitis.
Jejunum and Ileum.-The extreme rarity of actinomycosis in these regions has been explained by the rapid passage of the small-bowel contents. A more important reason may be the infrequency of penetrating lesions of the mucosa. A feature of the infection is the relatively normal appearance of the bowel at operation, the bulk of the inflammation occurring in the mesentery. Rigler (1964) described a case occurring in the ileum following perforation by a fish bone, and stressed the importance of complete excision where possible.
Ileocaecal Region.-This was the commonest site in the present series, as was also found in other reports (Cope, 1938). T h e source of infection in virtually all the cases is the appendix, ' periappendicular actinomycosis' being a more accurate name. T h e most interesting feature is the differing modes of presentation, which explain why this well-recognized condition is so often misdiagnosed. Cope (~9 3 8 ) classified pericaecal actinomycosis into two clinical types: (I) those which start with an attack of acute, usually perforated appendicitis; (2) those which start insidiously and first reveal themselves by the presence of a lump. He considered the former to be the commonest mode of presentation and certainly most of the cases described in the literature have come to light following an acute appendicitis (Wilson, 1961 ;Duncan, 1965). It is interesting that none of the cases in our series was of this type. Four presented as a mass in the right iliac fossa and I as recurrent attacks of small-bowel obstruction. After the disease has become established in this region, unless it is treated, it progresses in a classic way irrespective of its mode of presentation, with the formation of a hard craggy mass with associated sinuses or fistulae.
Colon.-Apart from the caecal region colonic actinomycosis occurs most frequently in the transverse and sigmoid regions. Because of the rarity of the disease in these sites the diagnosis is usually made late, a presumptive diagnosis of carcinoma, diverticular mass, or Crohn's disease usually being made. Palmisano and Russin (1969) and Salmo, Rudolf, and Makki (1969) both described cases occurring in the transverse colon. I n the latter case a preoperative diagnosis of carcinoma of the stomach was made; it was only after resection of the affected colon that the diagnosis came to light. Actinomycosis of the sigmoid colon is even rarer, but previous cases have been described by Cope (1938) and Piper, Stoner, Mitra, and Talerman (1969). T h e rarity in this region is difficult to understand as perforation of the bowel secondary to diverticular disease and carcinoma occurs most commonly in the sigmoid colon. An even more difficult fact to explain is the rarity of the infection following perforation of diverticula anywhere in the colon. We were unable to find any cases of actinomycosis following perforated diverticulitis. Although no cases of vesicocolic fistula have been recorded complicating actinomycosis, Piper and others (1969) described a case of ileovesical fistula which they attribute to actinomycosis arising in the sigmoid colon.
Anorectal Region.-Actinomycosis may arise primarily in the rectum and anal canal or may involve the perirectal tissues secondarily by downward spread from the iliac fossa. As would be expected, the infection tends to flourish in the loose tissues of the ischiorectal fossa, with a tendency to spread posteriorly rather than anteriorly (Morson, 1961). Perirectal actinomycosis may, therefore, come to light as an abscess in the buttock which recurs after drainage. A further progression of this local spread may explain the sacral abscess in the present Case 6 following actinomycosis in the right iliac fossa.
A special type of anorectal actinomycosis which took the form of an indurated ring at the anorectal junction was described by Morson (1961). He suggested that all rectal strictures should be biopsied to exclude this infection. Finally, a case of primary actinomycosis of the rectum was described by Fry, Martin, Dearing, and Culp (1965) in association with multiple perirectal and perianal fistulae, features which would lead most clinicians to suspect Crohn's disease.
Diagnosis.-Actinomycosis is one of the great imitators in clinical practice, particularly when it occurs in the abdominal cavity. Whilst reviewing cases presented by previous authors one is impressed by the length of time between the onset of symptoms and the final diagnosis. Our series is no exception to the rule as the average time to diagnosis was 18 months, ranging from I week to 6 years. Although the modes of presentation were varied there were certain features which, if borne in mind, may lead one to suspect the condition. T h e majority of cases demonstrates the 'latent' period referred to earlier, the symptoms following weeks or months after an abdominal catastrophe, which often involves perforation of the gut, or abdominal surgery. This may help to distinguish the condition from carcinoma, diverticular disease, and Crohn's disease.
A persistent discharging sinus, particularly following surgery for a perforated viscus, should always suggest the diagnosis, but unfortunately bacteriological study of the discharge may well be negative in the first instance. Even when specifically sought the classic colonies are difficult to culture, particularly as secondary infection of the abscess with such organisms as E. coli and staphylococci is so common. If visible macroscopically the 'sulphur granules' are pathognomonic but these are not always present, e.g., in only 5 of 13 cases at Manchester Royal Infirmary. Biopsy of the walls of sinuses or abscesses may also be fruitless, being reported on numerous occasions as showing ' non-specific ' inflammation.
Constitutional disturbances, although non-specific as far as diagnosis is concerned, may be particularly severe in actinomycosis. Pyrexia is common in the early stages but is often masked by antibiotic therapy. Microcytic anaemia is a constant feature. Although the erythrocyte sedimentation rate is often raised above ~o o m m . in the first hour, its value lies in monitoring the response to treatment rather than in diagnosis. Weight-loss usually occurs if the disease is widespread but again is not specific.
The diagnosis of actinomycosis rests almost entirely on the discovery of the typical colonies in pus obtained from the abscesses, but one must not scorn the pathologist for failing to see them initially if the possibility of their presence was not suggested by the clinician.
Treatment.-The first treatment for actinomycosis was surgical extirpation, and surgery still has a place today. Originally the aim was complete extirpation of the disease, which in most cases is impossible. Nowadays surgery is usually limited to incision and drainage with curettage of large abscesses. However, as recently as 1965 Duncan stressed the importance of an aggressive surgical approach in abdominal actinomycosis. Iodine was the first therapeutic agent to be used, and a palatable dosage of 5 minims in milk three times a day was suggested by Chitty (1926). It may still be a useful adjunct to the more usual therapy. The polyvalent actinomyces vaccine has been used as a specific therapeutic agent in the past. Placa, Proca, and Tomas (1966) revived its use in an abdominal case with success, following failure of antibiotic therapy. Its use nowadays lies in diagnosis in cases with negative biopsies. Following intracutaneous injection of 0.1 ml. of graded dilutions of the vaccine a local allergic reaction results in cases of actinomycosis (Bronner and Bronner, 1969). Radiotherapy has been used for its effect of softening fibrous tissue, but is not now used, mainly because of its effects on adjacent viscera.
With the advent of sulphonamides the prognosis of actinomycosis rapidly improved . Walker (1938) was one of the first to describe its use, suggesting a dosage of I g. three times a day. Since penicillin was first used no better treatment has been found. If penicillin is administered in the correct dosage it cures virtually every case. All the cases in our series were treated by penicillin with cures in all but I.
When first introduced, a daily dosage of up to 10 MU.
intramuscularly was recommended. This was very painful, required prolonged hospitalization, and has recently been shown to be quite unnecessary. A reasonable dosage is 2 MU. daily by intramuscular injection continued for 1-2 months after all clinical evidence of the infection has settled. In fact, our Case I was treated successfully with a considerably lower dosage, converting to oral penicillin in the later stages of the treatment. A feature of treatment that cannot be overstated is the necessity to continue the therapy for long periods after clinical extirpation of the infection. The cases showing a recrudescence of the disease after short-term therapy are legion.
Recently writers have described the use of other antibiotics in the treatment of abdominal actinomycosis with equally good results (Martin, Nichols, Wellman, and Weed, 1956;Svane, 1966;Halter, 1967). The use of more expensive and more toxic antibiotics appears to be unnecessary unless there is an allergy to penicillin or, as very rarely occurs, a resistant strain of fungus is encountered.