Lesson of the Week: Diabetic patients who do not have diabetes: investigation of register of diabetic patients in general practiceBMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6938.1225 (Published 07 May 1994) Cite this as: BMJ 1994;308:1225
The care of patients with non-insulin dependent diabetes mellitus is often considered the responsibility of general practitioners.*RF 1-3* Organising such care usually entails setting up a diabetic register, standardising care,4 and then performing an audit.5 While auditing our diabetic care we noted that some patients' serial glycated haemoglobin (HbA1) concentrations were all normal or nearly so. We decided to investigate whether these patients actually had diabetes.
Patients and methods
Our practice serves an isolated community of 8770 patients. A diabetic register was established in 1983. Although no formal criteria were used for entry, most patients had results from a blood glucose test (random, fasting, or postprandial) or an oral glucose tolerance test. A written management protocol was agreed in 1987; this included checking HbA1 concentrations in all patients every three months.
An audit of diabetic care was performed in October 1992. Patients with normal or near normal serial HbA1 concentrations were noted and their records examined. Those with no clear clinical evidence of diabetes and consistently normal serial concentrations of HbA1 had an oral glucose tolerance test in which blood glucose concentrations were measured after a 10 hour fast and then two hours after drinking 75 g anhydrous glucose. Two patients who were taking oral hypoglycaemic drugs had their treatment withdrawn one month before testing. Patients were classified as having normal glucose tolerance, impaired glucose tolerance, or diabetes mellitus according to World Health Organisation criteria (table I).6
Audit showed that we had 112 patients on our diabetic register. Twenty six patients had had normal or nearly normal HbA1 concentrations in the preceding six months. In 16 of these HbA1 concentrations had been high on previous occasions; the remaining 10 patients were selected for oral glucose challenge. Eight of them had had normal serial HbA1 concentrations (with occasional exceptions) over five years; the two others had had normal serial HbA1 concentrations but had become our patients comparatively recently.
Examination of the general practice records showed that none of the 10 patients had had the classic clinical symptoms of diabetes (thirst, polyuria, polydipsia, or pruritus vulvae) at the time of diagnosis. In one patient diabetes was confirmed after glucose challenge. Table II shows the results in the nine other patients.
Audits of diabetic care have focused on assessing process and outcome measures.5 An audit of referrals to a hospital outpatient department showed that the diagnosis was sometimes incorrect and poorly based.7 Our local biochemistry laboratory retains serial HbA1 results in all patients, and our practice is not alone in having a notable proportion of patients with consistently normal results (P Broughton, personal communication). In addition, HbA1 concentrations may be raised in conditions other than diabetes,8 so our register may include more non-diabetic patients. We found one such patient whose HbA1 concentration had never been checked because he did not believe himself to be diabetic. We suggest that the accuracy of diabetic registers should be investigated.
The coefficient of variation of blood glucose concentrations two hours after glucose challenge has been estimated at being between 20% and 35% over 12 months.9 We therefore advise follow up and repeat glucose challenge after a year in all patients repeatedly classified as having impaired glucose tolerance. The nature of impaired glucose tolerance has been debated; it clearly carries little risk of microvascular disease, though a few patients later develop diabetes.9
A diagnosis of diabetes has psychological and financial implications for patients. Added to this, three of our patients were exposed to the potential dangers of oral hypoglycaemic treatment. The primary health care team's resources and time are also implicated.
A diagnosis of diabetes requires care, particularly in the absence of clinical symptoms. A normal glucose tolerance may be adversely affected by starvation and other metabolic states, infections, gastric surgery, and some drugs, most commonly thiazides and steroids.10 Such common factors mean that a “normal range” for random blood glucose concentrations has little meaning, and values above the range are certainly not diagnostic of diabetes. Fasting estimations rely on patients' compliance. The formal oral glucose tolerance test has been criticised as “grossly overused in the clinical setting.”11 Nevertheless, we recommend that an oral glucose tolerance test be used to confirm or refute a suspected diagnosis of diabetes in patients without symptoms. Diagnostic thresholds depend on whether whole blood or plasma glucose concentrations are being measured.10
In recent years many practices have compiled registers for several chronic diseases. We suggest that other registers may be similarly flawed, as has been shown for registers of hypertensive patients.12 Protocols for management of all chronic diseases should include formal criteria for entry to an appropriate register.