The same was true for events attributed to
volume depletion. The explanation for these differences
is unclear, but could be the result of a publication bias, an
aspect of trial design, or diff erent characteristics of the
participants studied. Issues also probably exist with overreporting
of urinary tract infection, which was based on
symptoms alone in many cases. Notably absent from the
safety fi ndings is any indication of concerns about
acidosis and fracture,3–5 emphasising the time lag between
new safety issues being raised and trial data being
systematically reported. Findings from a programmewide
overview of acidosis reported for one SGLT2
inhibitor (canaglifl ozin) suggest a roughly three-times
increase in risk, but with only 12 events recorded,
substantial uncertainty remains about the scale of the risk
increase for this uncommon outcome.3 Notably, most
cases of acidosis occurred in individuals with an insulindependent
form of diabetes that had been misdiagnosed
as type 2 diabetes. Interim reports of a possible increased
risk of fracture have translated into a published scientifi c
report for only one drug (canaglifl ozin),72 and the extent to
which these eff ects might aff ect the class as a whole, or be
restricted to specifi c agents, remains unclear.