Tropical pyomyositis in the Solomon Islands: clinical and aetiological features

ROYALSocxn OF TROPICALMEDICINE AND HYGIENE(1989)

T~~~SA~~IONS

OF THE

Tropical features

pyomyositis

in the Solomon

83, 275-278

275

Islands: clinical and aetiological

R. Eason”, J. Osboume”, T. Ansford3, N. Stallman and J. R. L. Forsyth4 ‘Helena Goldie Hospital, Munda, Solomon Islands; %ntral Hospital, P.O. Box 349, Honiara; ‘State Health Laboramy, 63-69 George Street, Brisbane, Australia 4000; 4Microbiological Unit, University of Melbourne, Australia 3052

Abstract

A prospective study of tropical pyomyositis (TP) in the Solomon Islands’ Western Province followed 48 cases (mean age 10.6 years) from a population of 20 000 Melanesians over a two-year period. 32 patients were under 10 years and the male:female ratio was 1.7. Affected muscle was painful, swollen and often fluctuant but hard and indurated in presuppurative lesions. Abscesses, single in 40 and multiple in 8 subjects, were sited in the large muscles of the buttock, thigh, shoulder, arm and back on 52 occasions (90%). They contained 5-200 ml of pus which invariably grew penicillin-resistant Staphylococcus aureus sensitive to cloxacillin and erythromycin. All phage typable strains were identified as group 2 in contrast to the wider range of types found at carrier sites in otherwise healthy controls. TP was preceded by trauma in 30 cases (63%) and 26 (55%) of the patients had pre-existing pyoderma. Histological examination of clinically unaffected muscle biopsies from 10 subjects with solitary (7) or multiple (3) abscessesshowed no abnormality. Serological evidence of previous infection with adenoviruses or myxoviruses was present in the same proportion (4142%) of controls as of 22 patients tested. The results provide no evidence for antecedent diffuse myositis, viral or parasitic infections or nutritional deficiencies but support the role of trauma in localixing haematogenousskin staphylococci into damaged muscle. Introduction Tropical pyomyositis (TP) is characterized by iutramuscular abscesseswhich may be single or multiple and widely separated, and from which Staphylococcusaureus is the usual isolate. Whilst the pathogenesis remains unclear, prerequisites include skeletal muscle damage and staphylococcal bacteraemia. Trauma, infection and nutritional deficiencies have been suggested as agents of muscle damage which predispose to infection (EDITORIAL, 1978). This prospective study of TP was initiated to examine clinical and aetiological factors pertaining to its occurrence in the Solomon Islands.

(BLAIR & WILLIAMS, 1961).

Methods

The study area comprised the southern half of the Western Province of the Solomon Islands (latitude ‘Current address: 19 Colum Place, Bucklands Auckland,

8-Y south), where 20 000 Melanesian subsistence farmers, extsting on a diet of staple roots and tubers, green leaves, fruit and fish., live in the coastalregions. The subjects were all patients who presented to the Helena Goklie Hospital with intramuscular abscesses during the period July 1984-August 1986. In addition, two patients diagnosed clinically as having ileopsoas pyomyositis (low abdominal and hip pain; preference for hip flexion; tenderness above the anterior superior iliac spine; resolution after treatment with cloxacillin with or without surgical drainage (SCHECHTER,1983)) were included. Other presuppurative caseswere accepted only if supported by histological lindings in biopsies of affected muscle which were consistent with pyomyositis. Patients with subcutaneous abscessesor with muscle abscesses following penetrating wounds or pre-existing osteomyelitis were excluded. All subjects were questioned about the nature and duration of prodromal and presenting symptoms and about antecedent trauma and skin sepsis. Examination included a careful search for metastatic spread of infection and an assessment of nutritional status. Indices of wasting (weight for height) and stunting (height for age) were compared with sex-specific reference data collected by the United StatesNational Center for Health Statistics (WHO, 1983). Values indicative of malnutrition were delined as those corresponding to two standard deviations below the NCHS median. Fifty-two muscle abscessesfrom 42 patients were incised and the evacuated pus was examined locally with a Gram-stained smear while aliquots were flown to the laboratory, Honiara General Hospital for culture. Culture plates (blood and chocolatixed blood) were incubated aerobically at 37” for 48 h. Aerobic bacteria were identified by conventional methods. The antibiotic sensitivity of isolated organisms was determined by a disc diffusion technique (BAUER, 1966). During the second year of the study 11 coagulase-positive staphylococcal isolates were subcultured on nutrient agar and flown to the Microbiological Diagnostic Unit, University of Melbourne, for typing with the standard set of bacteriophages at routine test dilutions (RTD) and at 1OOxRTD

Beach,

New Zealand.

*‘Current address: 16 Alice Jackson Crescent, Gilmore, ACT. Australia 2905.

Staphylococci were also cultured from a control group of 18 patients with superficial abscesses.In addition, and in order to examine the strains of S. aureus prevalent in the catchment area, swabs from axillae, anterior nares and superficial skin lesions were taken ,from 104 residents of 4 villages, placed in :;~;~s transport mednun and examined in Mel-

276

Serological testing was performed on the first 22 patients seenand on 19 ageand sex matched controls. Blood was collected when patients presented and 2-4 weekslater. Serum was separatedand stored at - 10°C before being flown to the -Laboratory of Microbiology and Patholoav (State Health Laboratorv). Brisbane. Sera were e&ned for antibodies to &noviruses, myxoviruses, paramyxoviruses, cytomegalovirus and Toxdum c&ii. The results of comnlement &ation iests fo; antibodies to viral agent’s were interpreted as indicating recent infection when a four-fold rise in titre was noted in paired sera or a titre of at least 164 recorded in the initial specimen. Possible recent toxoplasmosis was diagnosed by a complement fixation titre of 1:8 and an immunofluorescence titre of 1:256. Evidence of infectious mononucleosis or endemic rickettsioses (Rickettsia typhi and R. tsutsugamushi) was sought by the Paul-Bunnell and We&Felix tests, respectively. Informed consent was obtained from 10 patients with pyomyositis for the taking of open biopsies of clinically normal gluteal or gastrocnemius muscle. These biopsies were taken during the second year of the study and were examined histologically at the Laboratory of Microbiology and Pathology, Brisbane. Full blood cotmt, chest X-ray and stool microscopy were performed on all cases. Results Clinical and pathological features

Forty-eight Melane& patients were admitted with TP durinn the w&d-an incidence of annroximately one caSeper 1000 population per ye&.‘The male:female ratio of the patients was 1.7 and their agesranged from 3 months to 63 years with a mean age of IO-6 years. 32 patients (67%) were under 10 years of age. The presentation of TP was characterized by pain and swelling of affected muscles for a mean of 8 (+5) d. Although 37 (77%) of cases were pyrexial (38-39°C) on admission, the patients were otherwise well and their constitutional symptoms mild. 42 patients had suppurative lesions and in half of these fluctuation was detected in the swollen tender muscle. Of the remaining cases, one was diagnosed as ileopsoas myositis, without demonstrable pus, which resolved with antibiotic treatment alone. In the other 5 pre-suppurative casesaffected muscles were characterized clinically by woody induration and histologically by florid infiltration with polymorphs, eosinophils and plasma cells and degeneration of muscle fibres. Single muscle involvement was noted in 40 subjects (83%) whilst 8 (17%) had multiple abscesses-7 cases with 2 foci and one with 4 foci of infection. The site of involvement was the large muscles of the buttock and thigh in 27 cases (47%), including two ileopsoas lesions (one suppurative and one pre-suppurative)? of the shoulder and arm in 17 (29%), and of the back m 7 cases(12%). The sternocleidomastoid was affected on 3 occasions,intercostal muscles twice and temporalis and soleus muscles once each. 31 lesions (57%) were on the left side and 27 on the right. Between 5 and 200 ml of pus were drained from 52 abscesses.In every casethe appearanceof clusters of Gram-positive cocci in the Gram-stained smear was corroborated by the culture of S. aureus. Only one

pre-suppurative biopsy specimen was cultured and it likewise yielded S. aureus. All isolates were resistant to penicillin and 15% were resistant to tetracycline. However, they were invariably sensitive to the other antibiotics tested-clindamycm., erythromycin and beta-lactamaseresistant penicilhns. 6 of 11 staphylococcal isolates from pyomyositic abscesseswere phage typable. All were identified as group 2 strains and lysed by phage type 71. Two were also sensitive to phages 55 and 3C. Additional tvping of stanhvlococci isolated from superficial abcesses-in 18 non-pyomyositic controls showed 10 of the 18 strains to be untvoable. 4 were identified as group 2 and 4 as group 3. In the survey for staphylococcal carriage in the village populations, swabs were taken from 104 villagers. Only 7% of axillary swabs, compared with 46% of nasal and 84% swabs from skin lesions, yielded S. uureus. From the 142 positive swabs, strains were selected for further study on the basisof differences in colonial morphology or pigment and haemolytic action. 492 strains (all methicillin sensitive) were submitted to phage typing. 351 (70%) were untypable at RTD or 1OOxRTD. Of the remaining 141, 12 were typed as group 1, 47 as group 2, 50 as group 3, 9 as group l/3 and 23 as miscellaneous. Such proportions were similar for all 3 sites. Mean packed cell volume and total white count was 38 (&4)% and 16.2~ 103/pl(f8*9x 103/& respectively. At 2.8 (*2*9)% and 0.4~ 103(+O*4x 103/pl), the relative and absolute numbers of eosinophils were within normal limits. No clinical or radiological features of chest infection were noted. After a 5 d course of cloxacillin and surgical drainage of muscle abscesses,complete healing and functional recovery was achieved after a mean of 13(f8) d. factors Thirty patients (63%) gave a definite history of significant trauma (25 falls, 3 kicks, 2 knocks) to the subsequently affected site. Trauma preceded the onset of muscle pain by an averageof 6( +4) d. In 3 of the 7 patients presenting with multiple abscessesthe first appeared at the injured site. Twenty-six (55%) of the patients noted pyodermata before the onset of pyomyositis. These lesions comprised minor trauma, impetigo following insect bites and furunculosis. While older children showed no signs of nutritional deficiency, 6 of 22 children under 5 years of age showed weight or height-for-age ratios, or both, indicative of malnutrition. However, this reflects the degree of malnutrition expected in this area and this age group (EASON, 1986). Serological evidence of adenovirus infection was present in 41% of 22 patients and 42% of controls. In only 2 patients and one control were titres indicative of recent infection. Other antibodies detected demonstrated past syncytial viral infection in 2 patients and past parainfhtenxa, past influenza and recent influenza in one patient each. Of the 11 patients who recalled having had a febrile prodromal illness (upper respiratory in 9 and diarrhoeal in 2), 8 were non-reactive to all serological tests performed. Direct examination of fresh faecesshowed helminth ova in 8 (17%) only (Ancylostumu in all; co-existing Contributiag

277 Trichuris in 2 and Ascaris in one) of TP patients compared with a prevalence of 25% for all stools examined in the hospital laboratory. The prevalence of previous toxoplasmosis was similar in patients (3 of 22) and controls (1 of 19). The histology of clinically uninvolved gastrocnemitts muscle from 7 patients with single and 3 with multiple abscessesshowed no evidence of diffuse myositis.

Discussion Male preponderance, age distribution, and clinical features are similar to those described for TP in most other reports (ADERELE & OSINUSI, 1980; AJAO & AJAO,

1982;

CHIEDOZI,

1979;

SIRINAVIN

&

MCCRACKEN,1979; SMITH et al., 1978). In all such studies, as in the current one, the most frequently affected areas were the large muscle groups of the buttock, thigh, shoulder and arm, which is probably just a reflection of their bulk. Multiple abscesses,reported in 15-40% of casesin other series (ADERELE & OSINUSI, 1980; AJAO & AJAO, 1982; CHIEDOZI, 1979) were present in 8 (17%) of our patients, one of whom presented with suppurative lesions of the right leg, left arm and left paraspinal muscles and- a further rhomboidal focus that manifested itself a week later. The occasional presenceof such widely separatedabscessessuggestsa transient bacteraemia as the primary event with local muscle insult resulting in focal seeding and bacterial proliferation (SCHECHTER, 1983; SIRINAVIN & MCCRACKEN,1979). Nutritional deficiencies (CHIEDOZI.1979).infection and trauma have been ‘implicated~ as the agents responsible for muscle damage. Indices of malnutrition in the current study were, however, restricted to 6 (27%) of 22 children under 5 years-a prevalence similar ’ to the 34% recorded in 1984 among 477 children of this age during an anthropometric survey within the studv area (EASON,1986). Previous reports of eosinophiha in casesof TP have led to the suggestion that migrating nematode larvae might causelesions predisposing to myositis (CHIEDOZI, 1979). However, such larvae characteristically travel between rather than within muscle tissues. Furthermore, our caseshad neither eosinophilia nor evidence of an undue frequency of parasitic infection. Taylor’s autopsy studies on casesof TP in Uganda (TAYLOR & TEMPLETON,1970) appeared to indicate the presenceof a generalized inflammatory in&ration in clinically unaffected skeletal muscle. This observation, combined with reports that TP may occur in close relatives and may be preceded by prodromal illness, has led to the suggestion that a viral agent may play an aetiological role (FANNEY& THOMAS,1982; SHEPHERD,1983). However, the histological data we obtained on biopsied unaffected muscles contrasted with Taylor’s findings. Furthermore, the lack of evidence of recent viral or rickettsial illness revealed by our serological examination supplements other studies (SHEPHERD,1983; TAYLOR & SHAW, 1973) which excluded a role for arbovirus or picornavirus. The skin is frequently the portal of entry for deep-seatedstaphylococcal infection and pyoderma, common throughout the tropics, was noticed by 55% of subjects before the onset of TP. S. aureus was cultured from over half such skin lesions during a

recent bacteriological survey in Papua New Guinea (BOWNESS& Bo-HER, 1984). When considerina the tvoable strains of S. aureus. all those isolated f$m pyomyositis belonged to phage group 2, whereas from nasal carriers and skin lesions among the villagers groups 2 and 3 strains were equally prevalent and group 1 strains not uncommon. The strains from superficial abscessesappeared to reflect the pattern of-those from skin lesions among villaaers. This is similar to Foster’s findings in UgGda 23 years ago (FOSTER,l%S). Howeve;, the very high proportion of untypable strains is very different and is more reminiscent of current methicillin-resistant strains. All strains in the current study were, however, methicillin-sensitive. The finding that 63% of our cases reported preceding traumatic injury to the affected site-and others could have forgotten an injury-is consistent with other studies and suggeststhat trauma could be an important precursor. However, minor injury is almost universal and could be coincidental and this factor doesnot explain the apparent freedom from TP of some tropical regions (SHEPHERD, 1983). By analogy with necrotiziug jejuuitis (pigbel) induced by clostridial exotoxin (LAWRENCE & WAL-

KER, 1976), SHEPHERD (1983) has recently suggested that a similar intestinal toxin might be the causeof the initial skeletal muscle injury in TP. The 2 countries with the highest incidence of plgbeduganda and Papua New Guinea-also have the greatest frequency of TP. The current study, however, negates this theory. While the incidence of TP within the study area of one case per 1000 population per year is as high asthat in New Guinea, only 2 casesof pigbel have been recorded here in the last 5 years (EASON& VAN RIJ, 1984)and protein intake in the study population is relatively high. Our data support the view that blood-borne skin staphylococci may localize in traumatized muscle to form abscesses,that pyodermata and trauma together are the principal determinants of such lesions, and there is no evidence to support a contributing role for malnutrition, viral or parasitic infection, or a diffuse myositis. The reported dissociation of trauma, infectious skin disease and pyomyositis in some other tropical regions (SHEPHERD,1983) remains a puzzle. Whilst absent from the present series, potentially fatal cardiac, pulmonary and cerebral suppurative complications have been reported (CHIEDO-ZL1979; SCRIMGEOUR & KAVAN, 1982) and orudence would dictate supplementing surgical drain~agewith a betalactamaseresistant penicillin or an alternative bactericial drug and follow-up review when possible. Acknowledgements

We acknowledgethe assistance of the Clinical Pathology Laboratory, Fairfield Hospital, Melbourne, in facilitating the transfer of strains of S. aureus and of swabs for examination and typing, and of Mrs Leonie Horvath, Mrs Rosemary Davidson and Mrs Jocelyn Hibberd for the phage-typing and for the examination of strains from the villages. This study was supported by a grant from the New Zealand Medical Research Council. References Aderele, W. I. & Osinusi, K. (1980). Pyomyositis in childhood. Joumal of Tropical Medicine and Hygiene, 83, 99-104.

278 Ajao, 0. G. & Ajao, A. 0. (1982). Tropical pyomyositis. International Sureerv. 67. 414-416. Bauer, A. W. (1966). Antibiotic susceptibility testing by a standardized single disk method. American Juumul of Clinical Pathology, 45, 493-$%.

Blair, J. E. & Williams, R. E. 0. (1961). Phage typing staphylococci. Bulktin of the World Health Grganization, 24, 771-778.

Bowness, P. & Bower, M. (1984). The bacteriology of skin sores in Gotoka children. Papua New Guinea Medical Journal, 27, 83-37. Chiedozi, L. C. (1979). Pyomyositis. Review of 205 casesin 112patients. AmericanJoumal of Surgery, 137,255-259. Eason, R. J. (1986). Childhood mahmtrition~in the western fr
27, 42-44.

Editorial, (1978). Tropical myositis. Lancer, i, 862. Fanney, D. & Thomas, L. C. (1982). An outbreak of pyomyositis in a large refugee, camp in Thailand. ;‘3E3~ Journal of Tropcal Medactue and Hygum, 31, Foster, W. D. (1965). The bacteriology of tropical myositis in Uganda. Journal of Hygiene, 63, 517-524. Lawrence, G. & Walker, P. D. (1976). Pathogenesis of enteritis necroticans in Papua New Guinea. Laucet, i,

125-126. Schechter, W. (1983). Tropical pyomyositis of the iliacus muscle in American Samoa. AmericanJoumal of Tropical Medicine aud Hygiene, 32, 809-811. Scrimgeour, E. M. & Kavan, J. (1982). Severe stauhvlococcaipneumonia complicating pyomyositis. Ame&nJournal of Tropical Medicine and Hygieue, 31, 822-828.

Shepherd, J. J. (1983). Tropical myositis: is it an entity and what is its cause? Lancet, ii, 1240-1242. Sirinavin, S. &McCracken, G. H. (1979). Primary suppurative myositis in children. AmesicanJoumal of Diseases in Children, 133, 26%265.

Smith, P. G., Pike, M. C., Taylor, E. & Taylor, J. F. (1978). The epidemiology of tropical myositis in the Mengo districts of Uganda. Transactions of the Royal Society of Tropical Medicine and Hygiene, 72, 46-53.

Taylor, J. F. & Templeton, A. C. (1970). Pyomyositis. A clinico-oatholoaical sNdv based on 19 autoosv cases. Mulago Hospi&i 1964-1968. East African Medical Jotaual. 147. 494-501.

Taylor, J. F. & Shaw, B. (1973). Tropical myositis: clinical and laboratory Studies. African 3ournal of Medical Science, 14, 409-%18. WHO (1983). Measuring change in nutritioual status. Geneva:

World Health Organization.

Received accepted

25 Jantuny

fat publication

1988; revised 16 August 16 August 1988

Eleventh seminar on amoebiasis 15-17 November

1989, Universidad National Aut6noma de Mexico, Ciudad Universitaria, M&tic0 City.

For further details residents of Latin America should contact Dr Librado Oritz-Oritz or Dr Morberto Trevirio, Instituto de Investigaciones Biomkdicas, Apartado Postal 70228, UNAM, Ciudad Universitaria CP 04510, Mexico DF; ail other persons should contact Dr Louis S. Diamond, Laboratory of Parasitic Diseases, Building 4, Room 126, National Institutes of Health, Bethesda, MD 20892, USA.

1988;