Background: Sepsis due to respiratory disease is one of the main complications of HIV/AlDS. Audit data from the Department of Internal Medicine at the University Teaching Hospital in Lusaka, Zambia indicate that pneumonia and tuberculosis in HIV represent two of the four leading causes of death. The mortality rate has remained high despite the advances made in antimicrobial spectrum. Limited ventilatory support options for patients with respiratory complications contribute to the high mortality rate. It is envisaged that non-invasive ventilation (NIV) will help reduce the mortality rate in this patient sub-population. Since Zambia has limited ICU capacity, less complex interventions such as NIV, might be lifesaving. Method: We conducted an observational prospective cohort study for the NIV arm (in the first half of 2016) with a retrospective chart review for the controls that focused on HIV positive patients with sepsis and hypoxaemic respiratory failure. 77 consecutive HIV positive patients with sepsis and respiratory distress meeting the inclusion criteria constituted the study population. Using the same clinical criteria 77 historical comparator patients were added from available charts reviewed from January 2014 to January 2016. After initial review of 385 folders, every 5^thfile was selected as comparator group. All patients meeting the inclusion criteria were offered NIV unless they opted out or refused to give consent. Clinical details were obtained and NIV initiated and clinical follow up recorded at 1, 24, 48 hours and daily assessment for 72 hours. Primary outcomes were patient tolerability on NIV. Secondary outcomes were survival to hospital discharge of participants on NIV. Clinical outcomes from the NIV arm were compared with the historical group. Results: 18 out of77 patients receiving NIV (23 .3%) died and the rest had a hospital survival to discharge. One patient in the NIV group left against medical advice after 48 hours, but his 24 and 48 hour clinical assessments showed marked improvement. In the historical group 64 out of 77 patients died by day 3, mortality rate of 83%. In both the historical and NIV groups, the main attributed cause of death was tuberculosis (89% vs 72% for the controls and NIV respectively). 4 patients developed complications of NIV (5.2%) leading to its discontinuation (3 patients had mask intolerance and worsening respiratory failure while one patient had gastric distension). Conclusion: Patients on NIV had a 72% relative risk reduction and a 60% absolute risk reduction of mortality compared to conventional group by day 3 at UTH. Only 5.2% developed intolerance or complication to NIV. Therefore, NIV could be a much needed arsenal in boosting survival outcomes in this patient subgroup.