Abstract 513

The predictive value of B and D BGs for subsequent BG levels has important implications for diabetes management. It has become common practice to prescribe acute adjustments in short acting insulin dosage based B and D glucose levels. However there is little quantitative data demonstrating the predictive relationship of B on L and D on S BGs to support clinical practice. We used archived records from patient BG meters (LifeScan) to evaluate the relationship of B on L, and D on S BG.

Patient records were uploaded at each clinic visit. Patients were taking fixed doses of human regular+NPH insulin once prior to breakfast and again before dinner. Patients followed their usual diet, activity and BG testing habits. Unique BG records for the B, L, D and S time slots were extracted for each patient and paired by the same date. Correlation analysis for a patient was performed if a patient's pairs exceeded 20.

There were 12,814 B+L pairs from 126 patients, average number of pairs was 96±83 (range 21-384) per patient, average correlation coefficient was 0.23±0.2 (range from -0.24 to 0.93). Ninety-five percent of the patients had correlation coefficients < 0.5.

There were 7528 D+S pairs from 85 patients, average number of pairs was 61±59 (range 21-346) per patient, average correlation coefficient was 0.25±0.2 (range -.24 to .85). Ninety-three percent of patients had correlation coefficients < 0.5.

The effect of ranges of B and D on percentage of immediately subsequent L and S readings >200mg% were tabulated. (Table)

Table 1 No caption available.
Full size table

Thus B and D have low correlation with respective L and S BG levels in the majority of pediatric patients with diabetes. This indicates the importance of other factors in determining the L and S BG. However the data also indicates that within broad ranges hyperglycemia >200 mg/dL occurs in the majority of subsequent L and S readings when preceding B and D glucoses exceed 200 mg/dL. Thus for patients receiving 2 daily insulin injections, upward adjustment of regular insulin might be a consideration when BG levels at B and D exceed 200 mg/dL. However acute insulin adjustments based on B and D blood glucose levels need to be carefully designed to insure safety and prevent increased variability in subsequent patient BG levels.