Sir,

We thank Pontiroli et al for their response to our article. As acknowledged in our article, a control group would have strengthened our study; however, a well-balanced or randomised design is difficult to achieve in bariatric surgery. Due to this limitation, we did not attempt to conclude a causal relationship between bariatric surgery and progression in diabetic retinopathy (DR). On the basis of our observations, we highlighted that contrary to common assumption, patients can continue to experience progression in their DR post-bariatric surgery.1 Therefore, it is of utmost importance that both patients and healthcare providers are aware of this potential risk and ensure patients attend regular DR screening.

We read with interest the recent cohort study that showed reduced incidence of microvascular complications in diabetic patients who experienced remission post-bariatric surgery.2 This is supportive of our findings, which showed that worsening of DR over time was more common in patients in whom glycaemic control failed to improve significantly postoperatively.1 Unfortunately, direct comparison with our study result is restricted by a number of factors—Coleman et al used ICD-9 diagnosis codes to define DR, which differs from the NSC grading system; detection of DR was dependent on patients seeking care rather than screening on all participants, and therefore there is a risk of misclassification as patients can experience asymptomatic deterioration in DR; follow-up was limited to 13 months compared to 4 years in our study. As highlighted by Mertlotti et al3 in their recent meta-analysis, based on available data from controlled studies—bariatric surgery may prevent new incident cases of DR, but currently available studies are not sufficient to support any systematic effect on progression or regression of DR. A number of limitations were noted—controlled studies were few, there was a low incidence of new cases of retinopathy and follow-up period was in general short. Therefore, as emphasised by them and many similar publications, further study is required to draw firm conclusions on the effect of bariatric surgery on DR, but progression of retinopathy can clearly still occur despite, in some cases, normalisation of glucose levels.