Intended for healthcare professionals


Comparison of blood or urine testing by patients with newly diagnosed non-insulin dependent diabetes: patient survey after randomised crossover trial

BMJ 1997; 315 doi: (Published 09 August 1997) Cite this as: BMJ 1997;315:348
  1. Pat Miles, diabetes nurse specialista,
  2. Joan Everett, diabetes nurse specialista,
  3. June Murphy, diabetes research nursea,
  4. David Kerr, consultant physiciana
  1. a Bournemouth Diabetes and Endocrine Centre, Royal Bournemouth Hospital, Bournemouth BH7 7DW
  1. Correspondence to: Dr Kerr
  • Accepted 27 February 1997


Guidelines for the management of non-insulin dependent diabetes mellitus often contain little information as to which is the preferred method of monitoring glucose at home.1 Although there is much enthusiasm for blood testing, previous studies have been inconclusive, probably reflecting the way patients were selected, as many have included people with poorly controlled, longstanding diabetes. Defects in patients' knowledge and diabetes education are likely to cause errors. Patients will also produce “dud results if their teachers are incompetent.” 2

We compared home testing of blood and urine in newly diagnosed patients with diabetes. All patients participated in an identical structured group education programme, beginning within a week of their diagnosis.

Subjects, methods, and results

In Bournemouth, patients with newly diagnosed diabetes are seen in groups of 6-16 in a nurse led group education session within a week of the diagnosis being made. Between December 1993 and December 1994 a total of 150 consecutive patients (91 men; average age 65 (range 31 to 91) years) were asked either to test once daily for glycosuria, alternating before or two hours after different meals, or at bedtime (target aglycosuria), or to test capillary blood glucose once daily before a different meal, or at bedtime, each day (target <8 mmol/l). The allocation was altered from week to week.

One month later, patients attended the first of four structured education sessions where all aspects of treatment, including starting drugs, were supervised by the diabetes specialist nurses according to predetermined protocols. After three months, patients were crossed over to the other method of home testing. Individual monitoring techniques were checked after one and four months. After six months, patients continued with their preferred method for a final six months.

Over the initial three months, 36 patients dropped out; 10 were unwilling or unable to manage blood testing, eight were unwilling to continue with urinalysis or had an altered renal threshold for glucose, nine were lost to follow up or died, two had poor vision, and seven changed their minds about the study.

Glycosylated haemoglobin (non-diabetic value <6.5%), body mass index (weight (kg)/height (m)2), and wellbeing were compared between groups (table 1).3 Patients compared methods for ease of use (70% (105) preferred urine testing and 15% (23) blood; 15% (22) were undecided), acceptability (44% (66) urine, 31% (47) blood), perceived accuracy (11% (17) urine, 76% (114) blood), and usefulness (21% (32) urine, 49% (74) blood).

Table 1

Comparison of home testing of blood or urine by 150 patients with newly diagnosed diabetes. Differences are with first measurement of variable

View this table:

At 6 months, 42% (63) opted for urine testing and 48% (72) for blood testing; 10% (15) chose both. Among the blood testers, 60% (43) were taking oral hypoglycaemics compared with 46% (29) of the urine testers.


For patients with longstanding diabetes, self monitoring of blood glucose does not necessarily improve glycaemic control nor facilitate weight reduction even when patients monitor accurately and compliance is excellent.4 Here, only patients with non-insulin dependent diabetes of recent onset were studied, and the education programme was identical for all participants. Over 12 months, glycosylated haemoglobin concentrations fell to the same extent, without any adverse effect on quality of life scores, with both methods of self monitoring. After six months, roughly equal numbers of patients chose each method.

In clinical practice, it is often difficult for patients with non-insulin dependent diabetes to maintain multiple daily tests over any length of time. Our data show that multiple testing may not be necessary: glycosylated haemoglobin concentrations at 6 and 12 months were similar to those achieved in the intensively treated group of the diabetes control and complications trial, in which subjects (young patients with insulin dependent diabetes) tested blood glucose four or five times a day.5

For patients with newly diagnosed diabetes who are able or willing to self monitor blood or urine for glucose, both methods are equally efficacious in terms of achieved glycaemic control, acceptability to patients, and self management. In economic terms, however, urine testing costs about one sixth as much as blood testing.


Funding: Urine and blood testing strips were kindly donated by Bayer Diagnostics.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract