From the article:
Professor Jones' team conducted a six year study of 587 men with type 2 diabetes, splitting them into three groups: those with normal total testosterone levels (above 10.4nmol/L, n=338), those with low testosterone levels (below 10.4nmol/L) that weren’t treated with testosterone replacement therapy (n=182), and those with low testosterone levels treated with testosterone replacement therapy for two years or more during the follow up period (n=58).
The findings show for the first time that low testosterone puts diabetic men at a significantly increased risk of death (p=0.001 log rank): 36 of the 182 diabetic men with untreated low testosterone died during the six year study, compared to 31 of the 338 men with normal testosterone levels (20% vs 9%). Furthermore, only 5 of the 58 diabetic men that were given testosterone replacement therapy died during the study (8.6%), meaning they showed significantly better survival compared to the non-treated group (p=0.049 log rank).