OBJECTIVES:  Many treatments and interventions are not aimed at treating a primary condition but rather at treating a secondary condition or symptom. Therefore, it is often desirable to have an understanding of the epidemiology and clinical burden of these secondary conditions. However the estimation of key epidemiological endpoints is often confounded by differences in study design and the definitions used for secondary conditions. Using the example of hyperglycaemia this research investigates how differences in study design and data reporting impact upon the estimation of prevalence and mortality. METHODS:  A search strategy was applied in Medline using a number of MeSH-based and free-text terms for “hyperglycaemia” and “intensive care units” for the period 2010–2016. A total of 10 studies were identified and selected that contained epidemiological data on both the prevalence and mortality of hyperglycaemia in the intensive care unit (ICU). RESULTS:  The 10 studies reported data from populations with differing primary conditions including: cardiac surgery, stroke, sepsis, diabetes, trauma, and myocardial infarction. The reported prevalence ranged from 10% to 87%, this large variation is most likely caused by differences in study design including cut-off level for hyperglycaemia (>140mg/dl in comparison to >200 mg/dl) and underlying condition. Similarly, heterogeneity was high for mortality (9–42%), this is likely to be due to differences in population characteristics (e.g. diabetes status and primary condition) as well as the time-point chosen for assessment of mortality. CONCLUSIONS:  This work highlights substantial heterogeneity across a small sample of epidemiological data for secondary conditions such as hyperglycaemia. A principle source of heterogeneity appears to be a lack of alignment in data reporting and study design.