Purpose: This study aimed to address 2 questions: First, what are the existing summary statistics of frequency of hypoglycemia in insulin-treated adults on established nasogastric feeding in the general ward? Second, to what extent does lack of homogeneity in defining, identifying, and reporting hypoglycemia affect these statistics?
Methods: A systematic review of the literature documenting hypoglycemia in insulin-treated adults on nasogastric feeding for ≥ 3 days in the general ward was carried out. Data sources were PubMed, Embase, ProQuest, Cochrane, Directory of Open Access Journals, and PLoS. Search period was 1999 onward, postdating introduction of analog insulin.
Results: Initially, 231 studies were identified, with 9 judged suitable for inclusion, according to inclusion/exclusion criteria. All included studies had as their primary objective the assessment of efficacy of insulin/feed regimens in the target population. Studies exhibited major heterogeneity. Definitions of hypoglycemia varied from < 60 mg/dL (3.3 mmol/L) to < 80 mg/dL (4.4 mmol/L), and 5 methods of reporting frequency of hypoglycemia were utilized, precluding pooled analysis. A descriptive synthesis of results was generated with some comparable results presented on a modified forest plot, showing 2.1% to 10.2% of patients with a hypoglycemic event and 1.1% to 5.4% blood glucose level < 70 mg/dL (3.9 mmol/L).
Conclusions: Hypoglycemia is not uncommon in this population, but further research is needed for quantification. Standardized documentation and reporting methods incorporating sample size and study duration, such as hypoglycemic events per patient-days, would facilitate interstudy comparisons, as would documentation of hypoglycemia at the 2 most commonly defined levels: < 63 mg/dL (3.5 mmol/L) and < 70 mg/dL (3.9 mmol/L).
© 2014 The Author(s).