Background: Nerve-function impairment (NFI) commonly occurs during or after
chemotherapy in leprosy and is the hey pathological process leading to dis
ability and handicap. We describe the development of a simple clinical pred
iction rule for estimating the risk of NFI occurrence.
Methods: New leprosy cases who presented to a centre in Bangladesh were rec
ruited and followed up for 2 years in a field setting. We used multivariabl
e regression analysis by Cox's proportional hazards model to identify predi
ctive variables for NFI. Discriminative ability was measured by a concordan
ce statistic. Internal validity was assessed with bootstrap resampling tech
niques.
Findings: 2510 patients were followed up for 2 years, 166 developed NF. A s
imple model was developed with leprosy group (either paucibacillary leprosy
[PB] or multibacillary leprosy [MB]) and the presence of any nerve-functio
n loss at registration as predictive variables. Patients with PB leprosy an
d no nerve-function loss had a 1.3% (95% CI 0.8-1.8%) risk of developing NF
I within 2 years of registration; patients with PB leprosy and nerve-functi
on loss, or patients with MB leprosy and no nerve-function loss had a 16.0%
(12-20%) risk; and patients with MB leprosy with nerve-function loss had a
65% (56-73%) risk.
Interpretation: Our prediction rule can be used to plan surveillance of new
leprosy patients. Patients at low risk of NFI may need no follow-up beyond
their course of chemotherapy (6 months); patients with intermediate risk n
eed a minimum of 1 year of surveillance; and patients with high risk should
have at least 2 years of surveillance for new NFI. Current recommendations
for surveillance of patients with leprosy (for the duration of chemotherap
y only) exclude an important group of patients who are at risk of developin
g NFI after completion of treatment.