OUP user menu

Glycaemic control in a type 1 diabetes clinic for younger adults

S.A. Saunders, M. Wallymahmed, I.A. MacFarlane
DOI: http://dx.doi.org/10.1093/qjmed/hch098 575-580 First published online: 18 August 2004


Background: In the Diabetes Control and Complications Trial (DCCT, 1993) tight diabetes control (HbA1c <7%) was associated with significantly less microvascular complications compared to ‘conventionally’ treated type 1 patients.

Aim: To assess the effectiveness of a dedicated young-adult type 1 diabetes clinic in achieving HbA1c levels <7% between 1991 and 2001.

Design: Retrospective review of case-notes.

Methods: All patients who attended the clinic in the first six years (1991 to 1996) were studied. All were offered at least two appointments per year. Case-notes were reviewed up to December 2001.

Results: We treated 386 type 1 patients (59.8% male, mean age 28.7 years, mean duration diabetes 9.5 years). After a mean follow-up of 7.7 years, 261 (67.6%) had attended in the past 2 years, 22 (5.8%) were known to have died, 11 (2.8%) had transferred to another clinic and 92 (23.8%) had repeatedly failed to attend appointments for 2 years. Over 11 years, the total mean (SD) HbA1c was 9.19% (1.3). Only 3.4% of patients achieved an average HbA1c of <7% during the study period, and 80% of patients had average HbA1c levels of >8%.

Discussion: Despite regular specialist physician, specialist diabetes nurse and dietician input, encouragement of multiple daily insulin injections and repeatedly following-up failed appointments (including home visits), fewer than 1:20 patients achieved the DCCT target of mean HbA1c <7%. Tight diabetes control is rare in a routine clinic setting.


The intensive therapy group of the Diabetes Control and Complications Trial (DCCT) demonstrated that tight glycaemic control (HbA1c <7%) in young adults with type 1 diabetes reduced the incidence of microvascular disease, when compared to conventional care at that time.1 The beneficial effects of tight glycaemic control in the DCCT on microvascular complications were maintained in the long term, despite subsequent deterioration in HbA1c values.2 Since the publication of the DCCT, tight glycaemic control, with HbA1c <7%, has been one of the central aims of diabetes care.

The intensive therapy arm of the DCCT was, however, expensive in terms of frequent clinic contact. Patients were reviewed on a monthly basis by a physician and had weekly contact with specialist nursing staff. Also, the rate of serious hypoglycaemia and the mean weight of patients in the ‘intensive therapy’ arm increased, something not seen in the ‘conventional therapy’ arm.

In 1991, a clinic for young adults with type 1 diabetes was established at our hospital. Since that time, it has been staffed by the same specialist diabetes physician and diabetes specialist nurse, with a dietician also present. Throughout, the main aims of the clinic have been optimal glycaemic control without troublesome hypoglycaemia, and screening for diabetic complications. This audit reports the glycaemic control achieved in this clinic and searched for factors associated with poor control.


The young-adult type 1 diabetes clinic was based in University Hospital Aintree, Liverpool, North-west England. The patients lived mainly in an urban area and originated from a wide spectrum of social groups.

The same diabetes specialist physician and diabetes specialist nurse have staffed the clinic during the study period. All patients saw the clinic dietician at least once and the great majority had multiple dietetic reviews. Follow-up appointments were offered at least twice-yearly, more problematic patients receiving more frequent follow-up appointments. Non-attenders were sent further appointments (at least three) and finally a home visit was made by the specialist nurse to encourage re-attendance and to maintain care.

There were 386 type 1 patients who attended the diabetes clinic in the 6 years between 1991 and 1996, and who had more than one recorded attendance. Data were obtained and reviewed from the case-notes of the clinic visits of these patients up until December 2001. At each attendance, HbA1c (DCCT-aligned) was measured, and total mean values throughout the study for HbA1c were calculated. The insulin dose (units/kg body weight) was recorded at last attendance, and compared with data from the end of the DCCT.2,,3 Attendance rates at diabetes out-patient appointments were calculated (attended/total appointments offered), and hospital admissions for ketoacidosis and hypoglycaemia recorded. Self-reported smoking habit at enrolment was noted, and known deaths were also recorded.

Statistical analyses

The individual patient data are not available for the DCCT patients. However, χ2 tests were used to compare prevalence data from the DCCT and Aintree patient groups. The mean values for each patient were combined to allow the calculation of total means, for the whole cohort. This was done for each variable. Data from the Aintree cohort are expressed as means (±SD). Comparison between groups within the Aintree cohort used the Mann-Whitney U-test, correlation was assessed by Pearson's correlation coefficient, and significance taken as p < 0.05.


Demographic data (Table 1)

The demographic data from the Aintree cohort are shown, compared to those of the DCCT cohort. The Aintree cohort contained more male patients (60%) compared to the whole DCCT cohort (53%) (p = 0.015, χ2). The age at first attendance of the Aintree clinic patients was similar to the age at which patients were recruited to the DCCT (mean ± SD age: Aintree 29 ± 10 vs. DCCT 27 ± 7). The DCCT subdivided the trial into two cohorts: one studying the appearance of microvascular disease in those with a short duration of diabetes and no evidence of microvascular disease (primary prevention group); and a second group, with a longer duration of diabetes in which the appearance or progression of microvascular disease was noted (secondary intervention group). The duration of diabetes in the Aintree cohort at recruitment was considerably greater: (9.5 years Aintree vs. 2.6 years DCCT primary prevention vs. 8.8 years DCCT secondary intervention). The patients in the primary prevention group of the DCCT were chosen specifically because they did not demonstrate any diabetic retinopathy, and had a duration of diabetes between 1 and 5 years. The Aintree cohort also had a significantly higher proportion of self-reported smokers (31%) than the total DCCT cohort (18%) (120/386 vs. 262/1441, p < 0.0001, χ2).

View this table:
Table 1

Demographic data at enrolment from the Aintree clinic and DCCT cohorts of patients with type 1 diabetes

Aintree patientsDCCT 1° preventionDCCT 2° prevention
n (males)386 (60%)378 (54%)348 (49%)352 (54%)363 (53%)
Age (years) at enrolment (mean ± SD)29 ± 1026 ± 827 ± 727 ± 727 ± 7
Duration of diabetes (years) at enrolment (mean ± SD)9.5 ± 8.12.6 ± 1.42.6 ± 1.48.6 ± 3.78.9 ± 3.8

Attendance and follow-up (Table 2)

Between 1991 and 1996, 386 patients attended the Aintree clinic who had more than one recorded attendance. At end of 2001, in the previous 2 years 261 (67.6%) had attended at least once, 92 (23.8%) had defaulted persistently, 11 (2.8%) were known to have moved away and 22 (5.8%) had died. There was a total of 4014 attendances during the study.

View this table:
Table 2

Glycaemic control (HbA1c) data from the Aintree Clinic Cohort up to December 2001

n (% total)Age (years) at diagnosis of diabetes (mean ± SD)Duration of diabetes (years) at enrolment (mean ± SD)HbA1c (%) (mean ± SD)
Attended in last 2 years261 (67.6%)18.6 (9.0)8.5 (7.1)9.1 (1.3)
Persistent defaulters in last 2 years92 (23.8%)19.6 (10.7)10.4 (8.8)9.2 (1.5)
Known moved away11 (2.8%)14.3 (7.6)9.9 (9.7)9.3 (1.5)
Died22 (5.8%)25.0 (11.8)*16.7 (9.6)*9.5 (1.5)
  • Mean follow-up 7.7 ± 3.01 yrs after enrolment. *p < 0.0001 vs. patients still attending, and persistent defaulters.

The Aintree cohort had a mean follow-up period longer than the patients in the DCCT (7.7 vs. 6.5 years). In the DCCT, 99% of patients completed the study and attended 95% of their hospital appointments; in contrast, 23.8% of the Aintree cohort had persistently failed to attend reviews in the last 2 years. Only 11 (0.7%) of the DCCT group died during the study period, compared to 22 (5.8%) of the Aintree cohort (p < 0.0001, χ2). The Aintree patients who died had significantly longer duration of diabetes, and were older at diagnosis of diabetes, compared to those who continued to attend the clinic and those who failed to attend in the past 2 years. The total mean HbA1c levels between these groups did not differ (Table 2).

The clinic non-attendance rate (defined as the ratio of appointments not attended to the total number of appointments offered during the study) was 0.32. There was a statistically significant, but weak, correlation between the non-attendance rate and higher mean HbA1c levels (r = 0.14, two-tailed p = 0.029).

Glycaemic control

The mean of all the HbA1c measurements over the study period from the 386 Aintree patients was 9.2%. This was very similar to the conventionally treated cohort of the DCCT (mean HbA1c 9.1%). Only 3.6% (14 patients) of the Aintree cohort achieved a total mean HbA1c <7% during the study. In comparison, the DCCT intensive therapy arm achieved a total mean HbA1c of 7.4% throughout the study.

Insulin administration and dosage

At the last recorded clinic attendance, the majority (58%) of the Aintree cohort were administering insulin by multiple daily injections (MDI—short- acting insulin, human or analogue, with meals, three times a day and isophane insulin at bedtime.) There was no significant difference in total mean HbA1c (%) between those patients administering insulin by MDI compared with those who elected to continue using human soluble and isophane insulin mixtures twice daily (9.1 ± 1.2 vs. 9.2 ± 1.3; p = 0.58). The mean amount of insulin (U/kg/day) at last recorded visit was not significantly different (0.75 ± 0.2 MDI vs. 0.67 ± 0.17 twice-daily insulin; p = 0.76). In the Aintree patients, the mean prescribed dose to all patients, of insulin per kg body weight at the last recorded patient visit was similar to that at the end of the DCCT: mean 0.74 U/kg/day (Aintree 2001, n = 386) vs. 0.75 U/kg/day (former DCCT intensive therapy group, n = 687) vs. 0.67 U/kg/day (former DCCT conventional therapy group, n = 688, data from EDIC enrolment2).


Severe hypoglycaemia, (requiring hospital admission) occurred in 0.79 per 100 patient-years of follow-up in the Aintree patients, similar to the conventional therapy arm of the DCCT (0.77 per 100 patient-years). The intensive therapy arm of the DCCT had a higher rate of admission, 1.14 per 100 patient-years (p < 0.001 vs. conventional therapy DCCT). Those Aintree patients who had severe hypoglycaemia (n = 18) had similar total mean HbA1c levels to those of the 368 patients who did not (9.3 ± 1.4 vs. 9.1 ± 1.3; p = 0.5).

Diabetic ketoacidosis (DKA)

In the Aintree cohort, the DKA rate was 2.39 episodes per 100 patient-years. This was higher than in the DCCT: 1.8 and 2.0 episodes per 100 patient-years in the conventional and intensive therapy groups, respectively. The 39 patients from the Aintree cohort who had least one admission with DKA had a significantly higher total mean HbA1c, compared to the 347 patients without such an admission (10.1 ± 1.1 vs. 9.0 ± 1.3; p < 0.0001).


The reductions in HbA1c levels achieved in the intensive therapy arm of the DCCT were associated with decreased rates in the appearance and progression of microvascular complications, in comparison to conventional therapy at that time. These differences in complications persisted, even though the differences in glycaemic control between the two therapy arms narrowed after the trial ended.4 The benefits of tight glycaemic control in preventing or delaying microvascular complications are therefore clear. But how can tight glycaemic control be achieved in routine clinical practice?

Our study directly compares the results from a young-adult type 1 diabetes clinic with the results achieved in the DCCT. The Aintree clinic, had from the outset, the aims of tight glycaemic control, i.e. the lowest HbA1c level possible, by encouragement of multiple daily insulin injections (four times daily), and they were reviewed by the same specialist doctor and nurse with dietician input. Freedom from troublesome hypoglycaemia, however, was also a major aim. The results show that, over a longer follow-up period than reported in the DCCT, glycaemic control was not as tight as the DCCT intensive therapy arm, and was similar to the DCCT conventional therapy arm. It is likely that the explanation for this is multifactorial. The Aintree patients had longer duration of diabetes, and unlike the DCCT patients, were not a selected group of highly motivated subjects who had enrolled in a trial. They also failed to attend one third of their clinic appointments, and the non-attendance rate was weakly associated with higher HbA1c levels. There was a loss of one third of the clinic population over the 11 years studied, from patients moving away, persistently failing to attend, or dying. Failure to attend UK diabetes clinics by 20–30% of patients has been documented previously.5–7 The total mean HbA1c of the persistent two-year non-attenders, was found to be similar to the total mean HbA1c of the patients who continued to attend the clinic. It is likely that many of these long-term defaulters had, in fact, moved away and failed to inform the clinic.

Compliance with insulin dosage is another factor influencing HbA1c levels. Although the recorded prescribed dose of insulin in the Aintree patients was similar to that given in the DCCT groups, it is likely that some Aintree patients did not comply. Evidence from Scotland suggests that up to 28% of young patients may not use insulin at the prescribed doses, leading to persistent under-use of insulin and chronically poor control.8

The avoidance of serious hypoglycaemia was an important part of the management of the Aintree patients, and hospital admissions with hypoglycaemia were fewer than in the intensive therapy group of the DCCT. Undoubtedly many Aintree patients would have reduced their insulin dose when hypoglycaemic episodes occurred, resulting in higher HbA1c levels in some patients.9

Omission of insulin by patients, for whatever reason, places them at risk of diabetic ketoacidosis (DKA). There was no apparent difference in the rates of admission to hospital as a result of DKA between the Aintree cohort and the DCCT patients, and no differences between the DCCT therapy arms. However, the Aintree patients who were admitted with DKA had a significantly higher HbA1c than those not admitted.

None of the Aintree patients were treated with continuous subcutaneous insulin infusions (CSII), in contrast to many patients in the intensive cohort of the DCCT. Some studies suggest that use of CSII may confer an advantage over MDI in achieving better glycaemic control.10,,11 However, there are obvious limitations of small studies using selected volunteer patients. Also, other studies have suggested there is no advantage to be gained in terms of glycaemic control by using CSSI.12,,13 CSII is more expensive than MDI administration, requires a highly motivated patient without psychological problems, and an experienced diabetes team who can provide regular and frequent input into the ongoing care of the patient.14

Recently, further strategies have been introduced to optimize glycaemic control. These include an intensive education programme; Dose Adjustment for Normal Eating (DAFNE),15 and newer analogue insulins.16–22 DAFNE may produce short-term improvements in HbA1c levels, but long-term data on sustained tight control is lacking. Again, a highly motivated patient willing to commit time and comply with advice is needed, along with considerable nurse educator resources.

The recently introduced long-acting analogue insulin glargine may be useful in improving fasting hyperglycaemia and reducing the incidence of hypoglycaemic episodes in patients with type 1 diabetes.17,19,20,23,,24 It is possible that the use of long-acting analogues may encourage the patient to aim for tighter glycaemic control, without the fear of hypoglycaemia leading to defensive reductions in insulin doses.

Is the goal of tight glycaemic control achievable in unselected type 1 clinic patients? The results from the Aintree cohort suggest that the great majority of patients will not achieve this, although long-acting analogue insulin and CSII were not used. Many barriers to tight glycaemic control exist in clinical practice that are not found in trial settings. Many patients do not comply with regimens long-term, fail to attend clinic regularly or move away from the clinic area. Appointments cannot be offered as frequently as visits in clinical trials, due to resource limitations. Despite the somewhat disappointing HbA1c levels achieved in this large cohort of young patients, well-organized, structured diabetes clinics have a very important role to play in the screening and early treatment of microvascular complications.


View Abstract