Glucocorticoid treatment and new-onset hyperglycaemia and diabetes in people living with chronic obstructive pulmonary disease: A systematic review and meta-analysis

Abstract

Introduction: In people living with chronic obstructive pulmonary disease (COPD), we aimed to estimate: (1) the prevalence of glucocorticoid-induced hyperglycaemia (GIH); (2) whether the prevalence of GIH varies by age, baseline diabetes status, treatment duration, ascertainment of glycaemia, definition of hyperglycaemia, study design and year of publication; and (3) the relative risk (RR) of new-onset hyperglycaemia in exposed vs non-exposed to systemic glucocorticoids.

Methods: We searched electronic databases until 9 November 2023 for randomised controlled trials and observational studies including adults diagnosed with COPD, with or without diabetes at baseline, using systemic glucocorticoids equivalent to prednisolone ≥5 mg/day for ≥3 days if exposed. Hyperglycaemia was defined as a blood glucose above a study-specific cut-off. We extracted data on study and participant characteristics, exposure and outcome. We performed random-effects meta-analysis to calculate pooled prevalence estimate of GIH. Prevalence was expressed as the proportion of people who developed hyperglycaemia among all exposed to systemic glucocorticoids during follow-up. We calculated RR of new-onset hyperglycaemia in exposed vs non-exposed to systemic glucocorticoids from eight studies.

Results: Of 25,806 citations, we included 18 studies comprising 3642 people of whom 3125 received systemic glucocorticoids and 1189 developed hyperglycaemia. Pooled prevalence of GIH was 38.6% (95%CI 29.9%-47.9%) with significant heterogeneity, I² = 96% (p < 0.010), which was partially explained by differences in study design. Pooled RR = 2.39 (95%CI 1.51-3.78). Publication bias was present.

Conclusion: The prevalence of GIH was 38.6%. Being treated with systemic glucocorticoids for COPD was associated with 2.4 times higher risk of new-onset hyperglycaemia versus no glucocorticoid treatment.

Keywords: clinical diabetes; diabetes; other complications; systematic review.

FIGURE 1

PRISMA flow diagram of the study.

FIGURE 2

(a) Prevalence of glucocorticoid‐induced hyperglycaemia in all 18 studies. (b) Prevalence of glucocorticoid‐induced hyperglycaemia in 13 studies which included people with and those without diabetes at baseline. (c) Prevalence of glucocorticoid‐induced hyperglycaemia in 5 studies, which included only people without diabetes at baseline.

FIGURE 3

Prevalence of glucocorticoid‐induced hyperglycaemia by subgroup. The p‐values are tests of significance for difference between the subgroup effects. In the definition of hyperglycaemia, category other included definitions based on fasting plasma glucose ≥5.5 mmol/L or ≥7.0 mmol/L or random plasma glucose ≥7.8 mmol/L. In the analysis of treatment duration, one study (McGraw et al. ⁴² ), one study was excluded, hence 17 rather than 18 studies. This was done because no details on treatment duration were reported in this study. In the analysis of the ascertainment of glycaemia, one study was excluded (Burt et al. ³⁴ ), hence 17 rather than 18 studies. This was done because this study used continuous glucose monitoring, a method not belonging to any of the categories. The analysis could be done with three or more studies per category.

FIGURE 4

Relative risk (95% CI) of new‐onset hyperglycaemia in people exposed to systemic glucocorticoids versus non‐exposed in 8 studies which reported data to calculate relative risk.