Statistical analysis for primary end-points:# type 1 and 2 diabetes primary analysis set
| Type 1 diabetes¶ | Type 2 diabetes | |||
| Within-subject difference+ | 95% CI of within-subject difference | Within-subject difference+ | 95% CI of within-subject difference | |
| Total cell count ×106 cells·mL−1 | 0.041±0.029 | -0.022–0.104 | 0.039±0.033 | -0.030–0.108 |
| Lymphocytes | 0.150±2.167 | -4.498–4.798 | -2.475±1.935 | -6.510–1.560 |
| Neutrophils | -2.070±1.449 | -5.178–1.038 | -1.788±0.834 | -3.527– -0.048§ |
| Macrophages | 1.840±3.095 | -4.798–8.478 | 5.125±2.113 | 0.718–9.532§ |
| Eosinophils | -0.846±0.606 | -2.109–0.417 | ||
Data are presented as per cent of total cell count, unless otherwise stated. CI: confidence interval. #: ANOVA with week and subject as fixed effects and an unstructured within-subject covariance matrix and the following covariates: centre, sex, atopic disease status and smoking history (never- versus ever-smoker). ¶: statistical analysis of eosinophils (percentage of total cells) was not possible because the data consisted mainly of values of 0.0. +: adjusted mean difference±sem between successive differences: (EXU-SC)-(SC-baseline). These differences were obtained through contrasts on the weekly values. §: statistically significant treatment group difference with an increase in percentage of neutrophils associated with SC insulin administration and an increase in percentage of macrophages associated with EXU therapy. SC: subcutaneous; EXU: Exubera®.