Peripheral neuropathy (PN) is associated with advanced HIV disease and may be a complication of antiretroviral therapy (ART) or anti-tuberculosis (TB) drugs, specifically isoniazid (INH). The effect of non-ART-drug-related PN on treatment outcomes is yet to be determined. We analysed prospectively collected cohort data for HIV-infected ART-naive adults initiating ART at the Themba Lethu Clinic, Johannesburg, South Africa from June 2004 to June 2009. Patients who presented with signs and symptoms of numbness or dysesthesia prior to initiation of ART were defined as having PN. Cox proportional hazard models were used to estimate the effect of PN alone (HIV-related PN) or PN with a history of INH use (TB-related PN) on mortality, lost to follow-up (LTFU), persistent and recurrent PN by 12 months of follow-up. Of the 9,399 patients initiating ART, 3.9 % had HIV-related PN while a further 1.8 % had TB-related PN. Patients with PN did not have a significantly higher risk of mortality compared to those without PN (hazard ratio (HR) 1.17 95 % CI 0.92-1.49). Patients with TB-related PN were less likely to be LTFU by 12 months (HR 0.65 95 % CI 0.44-0.97) compared to those without PN. Patients with HIV-related PN were at increased risk of persistent PN at 3 months post-ART initiation. Patients with HIV-related PN had a similar risk of recurrent PN compared to those with TB-related PN (HR 1.28 95 % CI 0.72-2.27). We demonstrate that patients with PN at initiation of ART present with advanced HIV disease. Completion of TB treatment may reduce the risk of persistent PN in patients with TB-related PN. Use of HIV drugs, even neurotoxic ones, may overall limit neuropathy.