The definition of HIV-SN as per the BPNS tool assesses both subjective and objective findings consistent with PN. However, the requirements that a patient must have both neuropathic symptoms and clinical signs to be diagnosed as HIV-SN may exclude some patients with mild disease  indicating a substantial underestimation of PN . This tool which is easy to use, practical and adds less than five minutes to the clinical examination of HIV patients during follow up in busy outpatient clinics  has been used in its standardised form in few studies. Using this tool, the prevalence of HIV-SN in our study population was 21% though, the prevalence of HIV-SN varies greatly among studies between 19 and 42% [16, 18–22]. According to Skopelitis and collaborators who used a model which involved both clinical criteria (symptoms and signs of neuropathy) and electrophysiology studies, two thirds of HIV-SN are subclinical . This therefore means that some patients though asymptomatic might have abnormal signs relevant to HIV-SN but which do not meet up with HIV-SN diagnostic criteria as per the BPNS tool. This was the case in 19.7% (58/295) of our study population. Some authors [19, 20] considered this group as asymptomatic PN who perhaps represent those with early HIV-SN who are more likely to become symptomatic when challenged with HAART or other risks for PN . Considering together those with HIV-SN (according to BPNS criteria) as symptomatic PN patients and those with clinical abnormalities relevant with PN but no symptoms as asymptomatic PN patients, increases the prevalence of PN in our study to 40.7%. In a resource limited setting like ours, this approach increases the likelihood of detection of patients with or at risk of PN. Consequently, caregiver choice of drugs of the HAART regimen would take into consideration each patient’s risk of PN and patient education on the avoidable risk factors would be accentuated. Either way HIV-SN is common among patients with HIV in Cameroon.
We found that HIV-SN is more common among older patients, similar with findings in other studies [4, 16, 19–21, 24]. This is because peripheral nerves by their length and size are known to have increased vulnerability with aging due to continual metabolic stress and exposure to toxic substances of which alcohol is among the most common as could be seen in our study where after adjusting for age, sex and height alcohol intake still remained strongly associated with HIV-SN consistent with findings in similar settings like ours . Morgello and collaborators in America also found that HIV-SN correlated with alcohol intake. Furthermore, increased life expectancy in the HAART era predisposes to longer exposure to the virus and/or to dideoxynucleoside analogues . Even though, there is considerable evidence on the risk of PN in the HAART era with use of neurotoxic agents [4, 17, 25], our study failed to show any such association especially with D4T. Another study  found that pre-existing PN before starting HAART reduced the risk that a patient would be placed on a D4T based regimen  thereby iterating the importance of pre-HAART screening for PN. However, the absence of association between D4T use and PN in our study could have been because only 15.6% of our patients were on a D4T containing regimen compared to as much as 98% in other studies in similar limited resource settings [19, 21].
A history of anti-tuberculosis treatment was found to be strongly associated to HIV-SN in our study. Isoniazid which is a major component of antiTB treatment during the whole treatment period has been shown to be a risk factor of PN .
Finally, our finding of low CD4 count, a reflection of advanced HIV disease, to be strongly associated with HIV-SN, was in line with findings in many other studies [4, 16, 17, 19, 21, 24]. Moreover, given that 83.9% of our symptomatic patients had symptoms before HAART initiation could be evidence that HIV-SN primarily might have been be due to the virus (HIV-DSP) .
Our study had some limitations. Firstly being a hospital based study in a tertiary institution where care is expensive; there is possibility of selection bias not permitting us to capture the real picture of HIV-SN among HIV patients in Cameroon. Secondly, a longitudinal design, unlike our cross-sectional design would enable us determine the incidence of ATN most especially as HAART coverage is being scaled up.