Assesment of the effectiveness of metformin in preventing prednisone-induced hyperglycemia among hematological cancer patients at Kenyatta National Hospital.

Abstract

Background: Steroid-induced hyperglycaemia is a common side effect of prednisone therapy among cancer patients. The prevalence, pattern, effects, monitoring, and treatment have been described in studies. Few studies have demonstrated the beneficial effect of metformin in preventing hyperglycaemia. Metformin’s preventative effect on prednisone-induced hyperglycemia in haematological cancer patients was investigated in low resource settings. Objectives: The overall goal of the study was to assess the effectiveness of metformin in the prevention of prednisone-induced hyperglycaemia among haematological cancer patients at Kenyatta National Hospital. Methods: A prospective randomized controlled trial of 24 cancer patients on high-dose prednisone was carried out at the Kenyatta National Hospital for 4 weeks. Eligible patients were randomized to either the intervention group receiving standard care plus metformin 850mg once daily for two weeks followed by 850 mg twice daily for another two weeks or the control group receiving the standard care. All participants had random baseline glucose levels were determined. The primary outcome was the presence or absence of prednisone-induced hyperglycaemia and patients had their fasting and 2-hour postprandial blood glucose monitored once weekly for the 4 weeks. All the analysis was done using STATA software version 13.0. Analysis of data was done using modified intention to treat analysis. The primary outcome, the presence or absence of prednisone-induced hyperglycaemia was dichotomized and expressed as proportions, with comparison across groups done using Fishers’ exact test. Comparison of inter-group variability of mean blood glucose differences was done using Mann-Whitney U test while within-patient comparison of single and double dose metformin within the treatment group utilized the Wilcoxon-Signed Rank test. Absolute and relative risk reductions alongside odds ratio was computed and reported along with their 95% confidence intervals estimates. A p value of <0.05 was considered as statistically significant. Logistic regression was employed to assess for the association between predictor variables and prednisone-induced hyperglycaemia. Results: Eighteen of the 24 randomized patients completed the study (11 control and 7 treatment). The proportion of the control subjects that progressed to pre-diabetes using the fasting and 2-hour postprandial glucose estimates was 72.7% (95% CI 45.5-90.9%) and 54.5% (95% CI 27.3-81.8%) xviii respectively. In contrast, the proportion of the treatment group was 14.3% (95% CI 0-42.9%) using fasting glucose, with no pre-diabetes being detected using the 2-hour postprandial glucose estimate. Comparative analysis of the mean fasting glucose in the 2 arms found no significant difference. However, statistically significant differences in mean 2-hour postprandial glucose in the 2 arms were noted in week 2 (p=0.0144), week 3 (p=0.0095) and week 4 (p=0.0074) of the study, where the treatment group presented with lower mean glucose values. Double dose (1700mg) metformin was more effective in lowering blood glucose than single dose(850mg), though this was not statistically significant using both fasting and 2-hour postprandial glucose (p=1.0000 and p=0.4531 respectively). Conclusion: The effectiveness of metformin in reducing the risk of prednisone-induced hyperglycaemia was significant. A prospective long-term study with a larger sample size can be employed to more conclusively elucidate the effectiveness of metformin in preventing steroid-induced hyperglycaemia.