Aims: The aim was to compare the use of four different systems of foot ulcer classification in a consecutive population with diabetes presenting to a specialist clinic in Dar es Salaam, Tanzania. Methods: Clinical data were collected prospectively in all patients presenting with foot ulcers between 3 January 2003 and 30 September 2005, and were used retrospectively to classify their ulcers using the Meggitt/Wagner, University of Texas (UT), Size (Area and Depth), Sepsis, Arteriopathy, and Denervation [S(AD)SAD] and Perfusion, Extent/size, Depth/tissue loss, Infection and Sensation (PEDIS) systems. Comparison was made between the strength of the associations between baseline characteristics of each system and outcome determined at 5 December 2005, using linear by linear association. Results: The strongest statistical associations (P < 0.001) were observed between percent healing and Wagner score (χ2 = 85.923), depth [S(AD)SAD, PEDIS and UT grade, 70.558], infection [S(AD)SAD, 61.774; PEDIS, 37.924] and UT stage (32.929). Weaker but significant (P < 0.001) associations were observed between percent healing and neuropathy [S(AD)SAD, PEDIS 12.475] and peripheral arterial disease [S(AD)SAD, PEDIS 10.799], as well as cross-sectional area [S(AD)SAD 4.387, P = 0.036]. Conclusion: The strength of the statistical association between outcome and both neuropathy and infection contrasts with findings in series previously reported from the USA and UK, and highlights the differences which may be found in different populations. These differences have implications for any system of classification chosen to compare the effectiveness of management in different centres in different countries. © 2008 The Authors.
