Gareth Hardy, HIV i-Base
Diabetes Mellitus (DM) occurs at an increased frequency in people with HIV infection and may develop at earlier stages than it does in the general population.
As antiretroviral therapy (ART) enables people with HIV to live longer, there has remained some controversy about whether or not the risk of diabetes is increased in HIV infection. 
Alfonso Hernandez-Romieu and colleagues at Emory University, Atlanta, Georgia, compared the frequency of diabetes mellitus in two large nationally representative health surveys, one for people with HIV and the other for the general population. They then examined the relationships between DM and other factors within and between these two groups. 
The Medical Monitoring Project contains data on 8,610 HIV positive people collected between 2009 and 2010. The National Health and Nutrition Examination Survey contains data on 5,604 adults in the general population. Logistic regression models were fitted to the data to determine the weighted prevalence of DM.
Using these models the researchers aimed to:
- Estimate the frequency of DM in this representative sample of HIV positive people;
- Compare the frequency of DM between HIV positive people and HIV uninfected people;
- Determine other factors that are associated with DM in HIV positive people.
Data was collected on socio-economic variables: age, sex, race/ethnicity, education, poverty level, as well as clinical variables: body mass index (BMI), time since HIV diagnosis, CD4 count, viral load, use of ART and HCV infection. The analysis did not include pregnant women and was restricted to adults over 20 years old and participants diagnosed with DM, defined as:
- Being told by a doctor that they had diabetes.
- Taking insulin.
- Taking diabetic medication to lower blood sugar (patients taking metformin monotherapy, who were not diagnosed with diabetes were excluded, as metformin can be used to treat pre-diabetes and other conditions).
The weighted prevalence of DM was calculated for HIV positive people and for each of the additional socio-economic and clinical factors. Multivariate logistic regression models were used to determine the relationships between DM and these additional factors.
The unadjusted frequency of DM in HIV positive people was 10.3% (CI 9.1-11.5%). This is higher than the US general population in which the frequency of DM was 8.3% (CI 7.2% – 9.4%). Among the HIV positive participants with DM, 3.9% (CI 2.9% – 5.2%) had type-1 diabetes, 52.3% (CI 46.7% – 57.8%) had type-2 diabetes and 43.9% (CI 38.1% – 49.8%) had unspecified diabetes.
Following an adjustment of the data for differences in the socio-economic variables, obesity and HCV infection, the adjusted prevalence difference in DM between HIV positive people and the general population was 3.8%. The largest difference in the frequency of DM in HIV positive people compared to the general population was in those with HCV infection (6.3%), those with high school or equivalent education (5.1%), women (5%), non-Hispanic whites (4.9%), those at or below the poverty line (4.6%), those with obesity (4.4%), and those aged 20-44 years (4.1%).
For people with HIV, the increase in adjusted prevalence of DM was lowest (6.7%) in younger people aged 20-44 years, whereas it was highest for those aged over 60 years (19.6%) and for those with obesity (18.9%). In HIV positive people without obesity the adjusted prevalence of DM was still increased at 7.9%. In addition, time since HIV diagnosis and geometric mean CD4 count were independently associated with DM in people with HIV. Interestingly, nadir CD4 count was not associated with increased prevalence of DM, in contrast to other studies. 
In conclusion, in this nationally representative US sample of HIV positive people, the frequency of diabetes was higher than the general population, and was independently associated with increasing age, obesity, longer time since HIV infection and CD4 count. After adjusting for other factors DM was 3.8% higher than in the general population. The data in this study suggests that HIV positive adults are at a higher risk of diabetes at a younger age and in the absence of obesity.
The researchers suggested that their results have important implications for care providers:
- Existing DM screening guidelines should be followed before and after starting ART.
- An examination of data from other prospective studies should be conducted to determine if screening guidelines need to be modified, as DM occurs at increased frequency in younger and non-obese HIV positive people.
- Improved tests for DM diagnosis and monitoring should be investigated in HIV positive people.
- Further research is required to establish optimal DM management approaches in HIV positive people.
There might be some different factors for US compared to European studies, but the conclusion to increase awareness of risk and to follow diabetes guidelines is important in all countries.
- Rasmussen LD et al. Risk of diabetes mellitus in persons with and without HIV: a Danish nationwide population-based cohort study. PLoS ONE 2012.7.12.
- Alfonso C Hernandez-Romieu at al. Is diabetes prevalence higher among HIV-infected individuals compared with the general population? Evidence from MMP and NHANES 2009-2010. BMJ Open Diabetes Research and Care. 2017. 5 (1). doi:10.1136/bmjdrc-2016-000304.
- Ghislain M et al. Late Antiretroviral Therapy (ART) initiation is associated with long-term persistence of systemic inflammation and metabolic abnormalities. PLoS ONE, 2015;10. e0144317.