Abstract
Background: Low glucose is a relatively common primary care referral to specialist endocrine services. However the prevalence of endocrine disease causing endogenous hypoglycemia is extremely rare. We sought to determine whether low plasma glucose in primary care was associated with adverse clinical outcomes or underlying hypoglycemic disorders. Methods: We obtained all plasma glucose results <72mg/dl originating from primary care within our catchment population (n = 800,00), in non-diabetic individuals (20,145 people [77.6% female]) aged 18 - 40 years, between 2002 and 2017. These data were linked to national admission, mortality, cancer and diabetes registers to assess associations with mortality and morbidity. Results: Median follow-up was 4.8 years (IQR 2.6 - 7.8). Median age was 29 years (24 - 34). Glucose was marginally higher in women 65 mg/dl vs. 63 mg/dl, P < 0.0001. Only 0.3% of patients were assessed in an endocrine clinic for evaluation of hypoglycaemia. Glucose concentration was < 40 mg/dl in 0.63% (category A), 40 - <54 mg/dl in 8.7% (category B), 54 - <63 mg/dl in 28.0% (category C) and ≥63 mg/dl in 62.6% (category D). A history of eating disorder was present in 2.4% (A), 1.1% (B), 0.4% (C) and 0.3% (D), P < 0.0001. Likelihood of admission to hospital with hypoglycaemia (P < 0.001) and electrolyte disturbance (P < 0.001) was also greater in those in the lower glucose categories. Increasing age (HR 1.03, P < 0.001) and male gender (HR 4.20, P < 0.001), but not glucose < 54 mg/dl (HR 0.89, P = 0.79), were associated with mortality (n = 72). The risk of a subsequent new diagnosis of cancer or hospital admission with incident cardiovascular, renal or liver disease was not related to glucose category. Hospital admission with infectious disease was more likely (3.3%) in those with glucose ≥ 54mg/dl compared to those with lower glucose (2.2%), P = 0.01. Incident diabetes was observed in 0.2% of those with glucose < 54 mg/dl and in 0.6% of those with glucose ≥ 54mg/dl (OR 3.0, P = 0.009). No cases of insulinoma were detected based on the results of these tests. Conclusion: Low plasma glucose results from primary care are almost never indicative of an endogenous hyperinsulinemic disorder and are not associated with adverse outcomes in adults up to 40 years of age. Underlying eating disorder should be considered in this context. Assessment by an endocrinologist should be limited to cases where Whipple’s triad is present.