Prognosis in tetanus: Use of data from therapeutic trials

Abstract

There is a need in clinical tetanus for a simple prognostic classification for all patients at time of admission to the hospital. Data from three randomized clinical trials performed in India in the 1960s, which contained information on several prognostic variables for 1,385 patients, have been used to study different methods of prognosis. A logistic regression performed on the combined data from the trials suggested that the probability of death is related separately to the period of onset (the time from the first symptom to generalized reflex spasms), to the time from first symptom to admission to the hospital, and to the clinical assessment of severity of tetanus on admission; the effects of these variables have been studied by tabulation. A new system is proposed for assigning patients to three prognostic groups, defined by time from first symptom to admission and whether or not reflex spasms were present on admission to the hospital. The range of fatality rates is 10%--63%. Neonates are assigned to a separate group with a fatality rate of 72%. Other prognostic groupings are also explored in case it is felt that the above method relies too heavily on local arrangements for medical care.