Post-surgical assessment of excised tissue from patients with Buruli ulcer disease: progression of infection in macroscopically healthy tissue

Authors


Authors
G. Bretzel (corresponding author) and V. Siegmund, Bernhard Nocht Institute for Tropical Medicine (BNITM), Hamburg; Department of Infectious Diseases and Tropical Medicine (DITM), University of Munich, Germany. Tel.: +49 8921 803618; Fax: +49 89 336112; E-mail: bretzel@lrz.uni-muenchen.de
P. Racz, F. van Vloten and B. Fleischer, Bernhard Nocht Institute for Tropical Medicine (BNITM), Hamburg, Germany.
F. Ngos, Hôpital de District, District de Santé d'Akonolinga, Akonolinga, Cameroon.
W. Thompson, Agogo Presbyterian Hospital, Agogo, Ghana.
P. Biason, Médecins sans Frontières (MSF), Geneva, Switzerland.
O. Adjei, Kumasi Centre for Collaborative Research in Tropical Medicine (KCCR), University of Science and Technology, Kumasi, Ghana.
J. Nitschke, Bernhard Nocht Institute for Tropical Medicine (BNITM), Hamburg, Germany; Médecins Sans Frontières (MSF), Geneva, Switzerland.

Summary

Objective  The current standard of treatment of Buruli ulcer disease (BUD) is surgical excision of lesions. Excision size is determined macroscopically assuming the complete removal of all infected tissue. However, dissemination of infection beyond the excision margins into apparently healthy tissue, possibly associated with recurrences, cannot be excluded in this way. To assess the central to peripheral progression of Mycobacterium ulcerans infection and the mycobacterial infiltration of excision margins, excised tissue was examined for signs of infection.

Methods  20 BUD lesions were excised in general anaesthesia including all necrotic and subcutaneous adipose tissue down to the fascia and at an average of 40 mm into the macroscopically unaffected tissue beyond the border of the lesion. Tissue samples were subjected to PCR and histopathology.

Results  Although the bacillary load decreased from central to peripheral, M. ulcerans infection was detected throughout all examined tissue specimens including the peripheral segments as well as excision margins of all patients. During the post-operative hospitalization period (averaging 2 months) no local recurrences were observed.

Conclusion  Available data suggest a correlation of surgical techniques with local recurrences. The results of this study indicate the unnoticed early progression of mycobacterial infection into macroscopically healthy tissue. Thus, the removal of all infected tissue cannot always be verified visually by the surgeon. Provided that long-term follow up of patients with positive excision margins will establish the clinical relevance of these findings, on-site laboratory assessment of excised tissue in combination with follow up may contribute to reduce recurrence rates.

Introduction

After tuberculosis and leprosy, Buruli ulcer disease (BUD) has become the third, in some West African countries even the second most or most common mycobacterial disease in immunocompetent humans (Amofah et al. 2002; Debacker et al. 2004). BUD occurs in tropical countries with foci in West Africa, Central Africa, and the Western Pacific. It is defined as an infectious disease involving mainly subcutaneous adipose tissue and the skin, characterized by a painless nodule, papule, plaque or oedema, evolving into a painless ulcer with undermined edges. It often leads to invalidating sequelae such as extensive scarring, contractures of joints with debilitating deformities and severe functional deficiency. Formal clinical trials to evaluate the use of anti-mycobacterial drugs have not yet been conducted, although anecdotal reports (WHO study group for drug treatment of Buruli ulcer, personal communication) suggest that some of these drugs used in combination with surgery or alone might have a beneficial effect. To date disease control in endemic countries is limited to early case detection through improved active surveillance and surgical excision of lesions. The current standard of treatment is surgical removal of all the affected tissue and part of the surrounding tissue, eventually followed by skin grafting (Meyers et al. 1996; Asiedu & Etuaful 1998; WHO 2000,2001; King et al. 2001).

Recurrence rates after different surgical techniques between 6.1% and 32% have been reported (Amofah et al. 1998; Kanga et al. 2003; Debacker et al. 2005). Teelken et al. (2003) describe large differences in treatment outcome between two hospitals applying different surgical practices.

To date the extent of the surgical excision depends on the surgeon's clinical experience alone. Figure 1 shows an ulcer before treatment. The excision size as shown in Figure 2 was determined based on the assumption of complete removal of all infected tissue.

Figure 1.

Lesion before surgery (6 months old, diameter 90 mm).

Figure 2.

Lesion after surgery.

However, dissemination of infection beyond the excision margins into macroscopically healthy tissue cannot be excluded visually. In general, little is known about the dissemination of bacilli and immunopathological changes in BUD lesions. With the intention to develop standards for determining the surgical excision size a laboratory-based approach to assess the spread of infection in BUD lesions was established. A pilot study on the post-surgical assessment of excised tissue was carried out in collaboration with Médecins sans Frontières (MSF), the Hôpital de District, District de Santé d'Akonolinga, Akonolinga, Cameroon (November–December 2003) and the Agogo Presbyterian Hospital, Agogo, Ghana (January until August 2004).

Materials and methods

Ethical clearance and informed patient consent

Ethical clearance for the study was sought through the Ministry of Health, Cameroon and the Committee of Human Research Publication and Ethics, School of Medical Sciences, University of Science and Technology, Kumasi, Ghana. Informed patient consent was obtained before surgery.

Inclusion criteria for study patients

Patients were eligible for inclusion if they fulfilled the following criteria:

  • Laboratory confirmation of M. ulcerans disease by both, PCR (Stinear et al. 1999) and histopathology (Guarner et al. 2003). Due to the long generation time of M. ulcerans culture, results were not considered as inclusion criteria.
  • Duration of the disease not more than 6 months.
  • Non-ulcerative (nodules and plaques) and ulcerative forms (elevated margins, moderate localized oedema, absence of dermatosclerosis).
  • Diameter of the lesion less than 10 cm.

Selection of patients eligible for the study and specimen collection was performed by the surgeon.

Patients, operative techniques and specimen collection

Twenty patients seeking treatment of BUD during the study period meeting the inclusion criteria (female: n = 10, male: n = 10, median age 19 years (range: 7–60 years) were included in the study (Akonolinga n = 10, Agogo n = 10). The lesions (ulcers: n = 19, nodules: n = 1) were localized at the lower limbs (n = 8), upper limbs (n = 7), back of the foot (n = 2), back of the hand (n = 1), abdominal wall (n = 2). The clinical aspects of these patients did not indicate any concomitant diseases such as tuberculosis or AIDS. Due to ethical reasons screening for HIV and other conditions was not part of the study.

All operations were performed under general anaesthesia. A pneumatic tourniquet was applied at the extremities. After disinfection the intended incision line was marked by a sterile pen. Excisions including all necrotic and subcutaneous adipose tissue down to the fascia were performed into the macroscopically unaffected tissue beyond the border of the lesion and were directed from peripheral to central. The average diameter of the excised tissue between the border of the lesion and the peripheral excision margin was 40 mm (range: 20–60 mm).

Excised tissue was kept in sterile cotton gauze moistened with sterile sodium hydrochloride. Specimens were taken with a sterile scalpel, gradient sections were cut from peripheral to central to avoid contamination. All specimens comprised skin and subcutaneous adipose tissue down to the fascia, where the bacilli are present.

Hospitalization and follow up of study patients

All study patients with ulcerative lesions received skin transplants and were hospitalized for approximately 2 months until complete healing. During the hospitalization period no local recurrences were observed.

Laboratory confirmation and laboratory methods

To assess the eligibility of patients for the study diagnostic specimens from the edge of the ulcer including necrotic tissue and all levels of the skin and subcutaneous adipose tissue down to the fascia were subjected to PCR and histopathology.

PCR, culture and histopathology were performed according to standardized procedures (Stinear et al. 1999; WHO 2001; Guarner et al. 2003). PCR results are indicated as ‘positive’ or ‘negative’, inhibited reactions were excluded from analysis. The histopathological classification of specimens as definite cases was based on the presence of acid-fast bacilli (AFB) and/or characteristical histopathological features. The bacillary load was graded as follows: absent, mild (1–5 AFB seen with 40 × objective), moderate (≥ 6 AFB seen with 40 × objective), or marked (AFB seen with 20 × objective as clumps or colonies) (Guarner et al. 2003). When AFB were absent and histopathological features did not allow a definite diagnosis as BUD, specimens were considered suspect cases unless they had other diagnoses that could account for a clinical nodule or ulcer. According to Meyers (1995) the healing stage of Buruli ulcer was added as a third group to the classification. Specimens were cultured to obtain M. ulcerans isolates for further studies, however, due to the long generation time of M. ulcerans and the low sensitivity of cultures, the results are not reported in this study (WHO 2001).

Gradient specimens

To assess the progression of infection from the inner edge of the ulcer up to the outer margin of the excised lesion, 20 gradient specimens were taken from excised tissue as follows: three parallel tissue sections (maximum size 50 × 15 mm) were cut from peripheral to central and subjected to PCR, culture and histopathology. Each section was cut in segments (maximum size 10 × 5 mm) (Figure 3). According to the size of the lesion the number of segments per set of gradients ranged from 6 to 3 (six segments: n = 2, five segments: n = 6, four segments: n = 5, three segments: n = 7). Altogether 83 tissue segments were taken for analysis. The diameter of the tissue sections was measured (starting point nodule: centre of the lesion, starting point ulcerative lesion: edge of the ulcer) and documented.

Figure 3.

Gradient segments (after cutting turned to the side to display subcutaneous adipose tissue).

All specimens were placed in test tubes (Sarstedt, Nümbrecht, Germany) containing 4 ml 10% formol for histopathology, 700 μl Cell Lysis Solution (Puregene DNA Isolation Kit, Gentra Systems, Indianapolis, USA) for PCR, 5 ml transport medium (Dubos Broth base + PANTA + Dubos Broth medium albumin, Becton Dickinson Diagnostic Systems, Heidelberg, Germany) for culture.

Excision margins

To assess if the margins of the excised tissue were free of bacilli, 20 sets of margin specimens (each margin segment maximum size 10 × 5 mm) were taken from either selected sites of the margin (large lesions) or the entire marginal tissue (small excisions) and subjected to PCR (Figure 4).

Figure 4.

Margin segments (after cutting turned to the side to display subcutaneous adipose tissue.

Anatomical reasons and the quality of excised tissue respectively allowed the laboratory analysis of 16 sets of margins only. The samples of the remaining four patients were either lacking subcutaneous adipose tissue or the excision margins were frayed, thus making it impossible to collect adequate specimens. In view of anatomical reasons it was impossible to take three parallel sets of specimens for histopathology, culture and PCR in analogy with the gradient samples. Therefore, only PCR analysis of margin samples was carried out.

According to the size of the lesion and the quality of the excised tissue the number of margin segments per set of margin ranged from 2 to 25 (2 segments: n = 2, 3 segments: n = 1, 5 segments: n = 1, 11 segments: n = 1, 14 segments: n = 1, 17 segments: n = 1, 19 segments: n = 1, 20 segments: n = 7, 25 segments, n = 1). PCR specimens were placed in test tubes (Sarstedt, Nümbrecht, Germany) containing 700 μl Cell Lysis Solution (Puregene DNA Isolation Kit, Gentra Systems, Indianapolis, USA).

Results

Gradients

As indicated in Table 1 after laboratory analysis of 20 sets of gradients 16/20 (80%) sets of gradients showed a positive result in all segments in both PCR and histopathology. 3/20 (15%) sets of gradients were found partially positive in PCR analysis (2/3, 2/4, 2/4 PCR positive segments, respectively), and positive in all segments in histopathology analysis. 1/20 (5%) set of gradients tested partially positive in PCR analysis (3/5 segments) and partially positive in histopathology (4/5 segments). 0/20 (0%) sets of gradients were found PCR positive in all segments and only partially histopathology positive. The overall accordance rate of PCR and histopathology results was 93.9% (77/82 segments were positive in both tests).

Table 1.  Polymerase chain reaction (PCR) and histopathology results of gradient segments
 PatientSegment aSegment bSegment cSegment dSegment eSegment f
  1. Segment a, central segment; BU, Buruli ulcer; histo, histopathology (AFB and histopathological classification); AFB, acid fast bacilli; pos, positive; neg, negative; NA: not available.

PCRA/19PosPosPosPosPosNA
Histo.A/19AFB mild
Definite BU
AFB mild
Definite BU
AFB moderate
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
NA
PCRA/20PosPosPosPosPosNA
Histo.A/20AFB moderate
Definite BU
AFB mild
Definite BU
AFB absent Healing stageAFB mild
Definite BU
AFB mild
Definite BU
NA
PCRA/21PosPosPosPosPosPos
Histo.A/21AFB marked
Definite BU
AFB marked
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
AFB absent
Definite BU
AFB moderate
Definite BU
PCRA/22PosPosPosNANANA
Histo.A/22AFB marked
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
NANANA
PCRA/23PosPosPosNANANA
Histo.A/23AFB mild
Definite BU
AFB absent
Definite BU
AFB mild
Definite BU
NANANA
PCRA/24PosPosPosPosPosNA
Histo.A/24AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
NA
PCRA/30PosNegPosNegPosNA
Histo.A/30AFB mild
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
AFB absent Not BUAFB mild
Definite BU
NA
PCRA/33PosPosPosPosPosPos
Histo.A/33AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
PCRA/34PosPosNegNegNANA
Histo.A/34AFB moderate
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
NANA
PCRA/38PosNegPosNANANA
Histo.A/38AFB moderate
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
NANANA
PCRK1PosPosPosPosNANA
Histo.K1AFB absent
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
NANA
PCRK 2PosPosPosNANANA
Histo.K 2AFB marked
Definite BU
AFB marked
Definite BU
AFB moderate
Definite BU
NANANA
PCRK 3PosPosPosPosPosNA
Histo.K 3AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
NA
PCRK 4PosPosPosNANANA
Histo.K 4AFB absent Definite BUAFB absent
Definite BU
AFB mild
Definite BU
NANANA
PCRK 5PosPosPosPosNANA
Histo.K 5AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB mild
Definite BU
NANA
PCRK 7PosPosPosNANANA
HistoK 7AFB moderate
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
NANANA
PCRK 9PosPosPosPosNANA
Histo.K 9AFB marked
Definite BU
AFB marked
Definite BU
AFB moderate
Definite BU
AFB mild
Definite BU
NANA
PCRK 12PosPosPosPosPosNA
Histo.K 12AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
NA
PCRK 15PosPosNegNegNANA
Histo.K 15AFB moderate
Definite BU
AFB mild
Definite BU
AFB mild
Definite BU
AFB moderate
Definite BU
NANA
PCRK 18PosPosPosNANANA
Histo.K 18AFB marked
Definite BU
AFB marked
Definite BU
AFB marked
Definite BU
NANANA

The peripheral segment of all histopathology gradient sections was positive for BUD [detection of AFB in 20/20 (100%) peripheral segments]. In general, as shown in Table 1, the quantity of AFB as detected by histopathology decreased from the central (segment a) to the peripheral segment. PCR analysis detected M. ulcerans DNA in 17/20 (85%) peripheral segments.

PCR did not detect M. ulcerans DNA in 6/82 (7.3%) specimens with a positive histopathology result. In most cases (5/7) PCR negative segments correlate with absence or low numbers of AFB as seen in histopathological analysis. Only histopathology could provide a definite diagnosis of BUD in all cases.

Margins

After laboratory analysis of all 16 sets of margins the specimens could be grouped as follows:

  • 100% margin segments completely PCR positive (n = 10);
  • >50% of segments per set of margin PCR positive [14/20 (70 %) n = 1];
  • <50% of segments per set of margin PCR positive [total: n = 5: 1/6 (16%) n = 1; 7/19, 3 inhibited (37%) n = 1; 2/25, 2 inhibited (8%) n = 1; 1/20, 9 inhibited (5%) n = 1; 7/20, 3 inhibited (35%) n = 1)];
  • 100% margin segments completely PCR negative (n = 0).

Discussion

To date surgical excision with subsequent skin grafting is still considered the most promising form of treatment of BUD. Nevertheless, despite broad acceptance surgical treatment has been discussed controversially. Consensus has been reached that wide surgical excisions of BUD lesions (i.e. comprising all necrotic tissue and the subcutaneous adipose tissue down to the fascia as well as macroscopically unaffected tissue beyond the peripheral border of the lesion) are to be performed (Cornet et al. 1992; Josse et al. 1995; Aguiar & Stenou 1997; van der Werf et al. 1999; WHO 2000).

However, standards for the extent of excision into macroscopically healthy tissue do not exist. Therefore, the surgeon's subjective experience alone determines the size of excision margins.

Only a few publications deal with surgical techniques (Ouattara et al. 2002, 2003, 2004; Knipper et al. 2004). Three case–control studies on recurrence rates and non-healed wounds after excision and skin grafting have been published: Amofah et al. (1998) report on 50 pre-ulcerative lesions from Ghana with a recurrence rate of 16% during a 1 year follow-up. Excision was performed under local anaesthesia with subsequent primary suturing. Recurrences are attributed to secondary bacterial infections and the surgical technique applied. Teelken et al. (2003) compare the surgical outcome after excisions of ulcers in two hospitals in Ghana and found non-healed ulcers (retrospectively not distinguishable from recurrences) in 18% (wide excision) and 47% respectively (excision close to border of lesion) of the cases subjected to follow up.

Kanga et al. (2003) report on 346 BUD cases from three hospitals in Ivory Coast with recurrence rates of 32%, 25%, and 11% (median 17.1%), within 1 year after excision. The low recurrence rate of 11% is attributed to the professional expertise of the surgeons. Kanga et al. discuss if a mycobacterial infiltration into the apparently healthy tissue of the excision margins could explain local recurrences.

The available data suggest two crucial issues for the patient outcome:

  • Painless surgery and wide excision, especially in children, are possible only with general, spinal or regional anaesthesia. Local anaesthesia, as often performed in pre-ulcerative cases, can make it difficult to achieve sufficient excision sizes. Moreover, local injection of liquid anaesthetic drugs may further promote peripheral diffusion of AFB into the adjacent subcutaneous tissue. Furthermore, as a general rule, healing of a septic process by secondary intention is achieved by laying the wounds open for later skin grafting or secondary suturing. Thus, local anaesthesia followed by primary suturing may favour recurrences as suggested by Amofah et al. (1998).
  • If a ‘wide’ excision into macroscopically intact tissue beyond the border of the lesion is performed, a mycobacterial infiltration into subcutaneous adipose tissue cannot be excluded visually. An excision size that seems clinically sufficient from the surgical point of view may therefore, due to macroscopically invisible progression of infection, be insufficient to be curative.

In order to assess the progress of infection from the centre to the periphery of the lesion and to examine the excision margins for signs of infection excised tissue from 20 patients from two hospitals in Cameroon and Ghana was subjected to laboratory analysis. The study patients were operated under generalized anaesthesia followed by skin grafting, thus local anaesthesia and primary suturing as possible risk factors for local recurrence could be excluded.

Although according to the healthy aspect of the excision line the presence of bacilli was not to be expected, signs of infection were detected in all segments of all sets of gradient specimens. Furthermore, in the majority of cases PCR rendered positive results for the complete set of margin segments, indicating the presence of M. ulcerans infection around the entire excision margins.

These findings indicate that the ‘wide’ excision into macroscopically healthy tissue that has been performed in all cases and was assumed to be curative from the surgical point of view, in fact was not sufficient to remove all infected tissue. The macroscopic aspect of apparently healthy tissue does not correlate with the dissemination of infection into subcutaneous adipose tissue. Thus, a reliable determination of excision margins sufficient to remove all infected tissue based on macroscopic criteria alone is not possible.

In this study the average excision diameter of 40 mm did not result in M. ulcerans free margins. In order to develop standards for the excision size, further studies on specimens from areas beyond the excision line are required. The clinical relevance of these findings is yet to be proved. Short-term follow up of the study patients during hospitalization for a period averaging 2 months until completion of skin grafting did not reveal any local recurrences. However, to establish a correlation between the presence of persisting M. ulcerans infection in excision margins and the risk of local recurrences a long-term follow up is required. Quantification of the bacterial load in margin segments according to the method described by Rondini et al. (2003) may support these investigations by elucidating the question if a minimal bacterial load is required for the development of recurrences and if the self-healing potential can be related to the bacterial load.

On-site assessment of excised tissue may provide an option to guarantee the complete removal of infected tissue. According to the method described by Siegmund et al. (2005) pre-surgical analysis of punch biopsies taken from the intended incision line could help identify the adequate excision size. Post-surgical assessment of excision margins as carried out in this study may contribute to identify patients with incompletely removed infected tissue. In this study the sensitivity of PCR was slightly lower than histopathological analysis, which could be explained by the fact that specimens from parallel tissue sections, and not one and the same specimen, were subjected to analysis with both methods. To gain the maximum diagnostic sensitivity, PCR negative specimens could be re-examined by histopathology, if this technique is available in the respective setting. Patients who tested positive must be subjected to a regular follow-up of at least 12 months or even to early re-operation in case of persisting wounds or recurrences. As a prerequisite to achieve such long-term follow up however, capacities for surveillance of patients in their villages must be made available. In addition, the possibility of supportive antimycobacterial treatment should be considered to reduce recurrence rates. Drug trials evaluating antimycobacterial therapy alone and in combination with surgery conducted by research groups from Belgium and the Netherlands are scheduled to commence in 2005 (Francoise Portaels, Tjip van der Werf, personal communication).

Acknowledgements

This work was partially supported by the Volkswagen Foundation (Grant number Az.I/79 314). The authors would like to thank Médecins sans Frontières, Geneva, for providing access to the patients treated at the MSF surgical project Akonolinga, Cameroon. We are grateful to Anja Schörle, Petra Meyer and Birgit Raschdorff for excellent technical assistance.

Ancillary