Background Sub-Saharan Africa suffers from a dual burden of infectious and non-communicable diseases. There is limited data on causes and trends of admission and death among patients on the medical wards. Understanding the major drivers of morbidity and mortality would help inform health systems improvements. We determined the causes and trends of admission and mortality among patients admitted to Mulago Hospital, Kampala, Uganda. Methods and results The medical record data base of patients admitted to Mulago Hospital adult medical wards from January 2011 to December 2014 were queried. A detailed history, physical examination and investigations were completed to confirm the diagnosis and identify comorbidities. Any histopathologic diagnoses were made by hematoxylin and eosin tissue staining. We identified the 10 commonest causes of hospitalization, and used Poisson regression to generate annual percentage change to describe the trends in causes of hospitalization. Survival was calculated from the date of admission to the date of death or date of discharge. Cox survival analysis was used to identify factors associate with in-hospital mortality. We used a statistical significance level of p<0.05. A total of 50,624 patients were hospitalized with a median age of 38 (range 13–122) years and 51.7% females. Majority of patients (72%) had an NCD condition as the primary reason for admission. Specific leading causes of morbidity were HIV/AIDS in 30% patients, hypertension in 14%, tuberculosis (TB) in 12%), non-TB pneumonia in11%) and heart failure in 9.3%. There was decline in the proportion of hospitalization due to malaria, TB and pneumonia with an annual percentage change (apc) of -20% to -6% (all p<0.03) with an increase in proportions of admissions due to chronic kidney disease, hypertension, stroke and cancer, with apc 13.4% to 24%(p<0.001). Overall, 8,637(17.1%) died during hospitalization with the highest case fatality rates from non-TB pneumonia (28.8%), TB (27.1%), stroke (26.8%), cancer (26.1%) and HIV/AIDS (25%). HIV-status, age above 50yrs and being male were associated with increased risk of death among patients with infections. Conclusion Admissions and case fatality rates for both infectious and non-infectious diseases were high, with declining trends in infectious diseases and a rising trend in NCDs. Health care systems in sub-Saharan region need to prepare to deal with dual burden of disease.

Commercially available rapid strip assays (RSAs) for hepatitis B surface antigen (HBsAg) are used for most routine clinical testing in sub-Saharan Africa. This study evaluated the validity of RSA and a more sophisticated enzyme immunoassay (EIA) with confirmation by nucleic acid testing (NAT) in hospitalized patients in Uganda. Sera from 380 consecutive patients collected and tested for HBsAg and anti-HIV in Kampala, Uganda by RSA were sent frozen to Dallas for EIA including HBsAg, total anti-hepatitis B core, hepatitis B e antigen, and anti-HIV. NAT was performed on all HBsAg-positives and on a random sample of 102 patients that were HBsAg-negative by both assays. Overall, 31 (8%) were HBsAg positive by RSA while 50 (13%) were HBsAg-positive by EIA; 26 were concordant between the two assays. Of 55 HBsAg-positive patients, nearly all showed detectable serum hepatitis B virus (HBV) DNA by bDNA (46) or PCR (4) assay. The 26 patients who were HBsAg positive by both EIA and RSA had significantly higher median serum HBV DNA levels than the 24 patients who were HBsAg positive by EIA alone. An additional 12/102 (12%) HBsAg negative patients had very low serum HBV DNA levels by NAT. Several differences in expected results of serologic testing were observed in this large series of African patients. RSA HBsAg testing is less sensitive than EIA; even EIA failed to detect all HBV DNA positive sera. A more complex testing protocol than RSA alone will be needed in Africa to improve patient care.

Most hepatitis C testing in Uganda is performed using commercial rapid strip assays (RSA) to detect antibodies to hepatitis C virus (anti-HCV), rather than enzyme immunoassays (EIA). The prevalence of hepatitis C antibodies in a Ugandan hospital population was determined using both methods to test their accuracy using nucleic acid testing (NAT) as a reference. Sera from 380 consecutive hospitalized Ugandan patients were tested for anti-HCV using an RSA in Uganda, with subsequent automated third-generation EIA testing in the United States, followed by NAT. Recombinant immunoblot assays (RIBA) were used as a supplementary test to detect anti-HCV epitopes. Overall, anti-HCV was detected in 48/380 (13%) by one or both antibody tests. Anti-HCV was detected in 19 (5.0%) patients by RSA and in 33 (8.7%) patients by EIA; only four patients were anti-HCV positive by both methods. Fourteen of the 48 anti-HCV positive patients had detectable serum HCV RNA, 7 each by bDNA assay or by PCR. RSA detected only 7 of 14 HCV RNA positive sera. Of 29 RNA negative but anti-HCV positive patients tested by RIBA, only two were anti-HCV positive; 27 were anti-HCV negative or indeterminate. Anti-HCV testing by RSA and/or EIA was neither sensitive nor specific for detection of ongoing HCV infection in hospitalized Ugandan patients. Our findings underscore the importance of confirmatory nucleic acid testing, which, despite its increased cost, appears essential to manage African patients with HCV.

Background: Hepatitis B virus (HBV) is the commonest cause of primary hepatocellular (PHC) carcinoma worldwide. Coinfection with the HIV leads to more rapid progression of liver disease. Objectives: We described prevalence of HBV and HIV among patients with PHC admitted to Mulago Hospital, Kampala, Uganda. Methods:We assessed all patients admitted to the gastrointestinal service of Mulago hospital with a diagnosis of PHC for HBV and HIV infection. Results: From March to June 2008, we recruited 15 patients. Nine (60%) were male; the overall median age was 32 years (IQR 15 -67), with median ages for male and female 33 and 36 years respectively. Alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase and AFP were all elevated with median values of 57.5 IU/L, 222 IU/L, 392 IU/L and 362 ng/ml respectively (IQR 14-145, 49-393, 165-1294 and 7-480). Eight (53%) patients were from North and Northeastern Uganda. The HBsAg was reactive in 13(87%) patients and HIV in 3(20%), all of whom were also co-infected with HBV. Conclusion: There is high prevalence of HBV and HBV/HIV co-infection among patients with PHC in Uganda with high mortality. Reduction in incidence and mortality due to PHC in Uganda will require urgent large scale HBV vaccination.

Background: Dyspepsia is defined as a chronic or recurrent pain or discomfort centered in the upper abdomen. Endoscopy is the best strategy for confirming the cause of dyspepsia. Non-invasive strategies would be more appropriate in low resource countries where endoscopy is not readily available. However, there is concern that these strategies may miss serious disease like gastric cancer. One test that needs to be assessed in this regard is the Helicobacter pylori stool antigen test (HPSAT). Objective: To determine the validity of the stool antigen test in predicting H. pylori associated disease among patients with dyspepsia. Methods: In this prospective study patients with dyspepsia attending Mulago Hospital were recruited consecutively. Helicobacter pylori was determined using the Rapid Strip HpSA ®, endoscopy and gastric mucosal biopsy were done. Results: 167 patients with dyspepsia were recruited into the study. There were ninety six (57.5%) females and seventy one (42.5%) males with an average age of 48.1(±18.1) years. Patients presenting with dyspepsia in Mulago hospital were more likely to come from the Central 60 (36%) and western tribes 55 (33%). The commonest endoscopic finding was oesophagitis 25 (15%). Peptic ulcer disease was found in 32 (19.2%) and 54 (32.3%) had normal endoscopy findings. H pylori was found in 33.5% and 32.5% using the HPSAT and histology respectively. The validity of the HPSAT in predicting H.pylori associated diseases was generally low with an overall sensitivity of 55.8%, and specificity of 74.2%. However, the validity was higher in predicting the diagnosis of peptic ulcer disease with a sensitivity 59.4% and specificity 72.6%. Conclusion and recommendations:The HPSAT may be used in the test and treat strategy for young patients with dyspepsia without alarm signs and symptoms in low resource settings. However, because of its low validity in predicting H.pylori associated disease, it is important to follow up patients so that if symptoms persist or recur endoscopy is performed