This is highly interesting report providing valuable information to the research community about the effectiveness of metreleptin substitution in patients with lipodystrophy in real-life setting.

Given the extreme rarity of all lipodystrophy sub-forms, a remarkable large group of patients is presented. The very detailed analysis is helpful in understanding the effect on the different parameters.

Comments on the different sections:

Methods
• Definition of responders vs non-responders:
o The definition used in this study for efficacy on diabetes (0.5-point decrease in HbA1c, or HbA1c stability with a decrease of more than 50% in total daily insulin, or discontinuation of at least one antidiabetic class has) been defined before by the same authors (Ref 21) – but not by Ref 24 (Oral et al) who suggest a decrease of HbA1c >= 1%

• Definition of hypertension
o In children below the age of 16 there is no fix cut-off like in adults but hypertension in children is defined as systolic BP (SBP) and/or diastolic BP (DBP) persistently ≥95th percentile for sex, age and height measured on at least three separate occasions. This may be relevant for some of the patients with GLD and should be commented on.

Statistics
• It is not mentioned whether or not the p value is two-sided. At least for comparison short term vs long term it should in fact be – so probably this would be appropriate for the whole analysis.

Results:
• Are truly 63% of PLD patients treated with insulin non-responders concerning glucose homeostasis? If all non-responders (n=7 according to Supplemental Figure 1) are within the insulin treated group (n=12 according to Table 3) that would be 7 out of 12 which makes only 58%. And what does the p value refer to?

• The comment that children with GLD gain weight, in keeping with growth development may be correct, but that is in fact not what Table 2 shows. Table 2 shows a mild increase in z score. But even children with no change in z score gain weight over time as the expected weight rises with age so the same z score translates to a different weight over the time.

• It would have been nice to learn about changes in pancreatic events – but probably the retrospective analysis was not providing this data?

• Could you comment on changes in rates of dyslipidemia, diabetes mellitus and hypertension?

Tables
• Table 2
o Weight z-score is not given in kg (please delete the kg in brackets)

• Supplemental Table 3,
o P values for patients treated with insulin are in the wrong line

o Differences in number of antidiabetic drugs – please provide more digits after point (right now it reads 1.0 is larger than 1.0 than 1.0)

o Albuminuria in text: albuminuria does not significantly change at long-term versus baseline (mean 354mg/L vs 345 mg/L, respectively p=0.94) – in Table: 33.5 mg/L vs 21.3 mg/L

• Table 2 and Table 3; Differential response in patients with GLD and PLD
o Number of antidiabetic drugs can not be correct (median for GL 1.0, median for PL 2.0, median for whole group 1.0????)

• Table 3 + Figure 2A
o HbA1c long term response 7.6 in Table 3 but seems to be 8.x in Figure 2A

Discussion
• Line 321.322 is unfortunately not correct: “Importantly, it confirms that metreleptin treatment is efficient, in a sustainable manner, in reducing hyperglycemia, hypertriglyceridemia and liver enzymes in patients with GLD.” As Table 2 shows, metreleptin does not lower liver enzymes over a longer time period.

• The observation that insulin treatment was a negative predictor for efficacy of metreleptin treatment in patients with PLD in line with previous examinations might benefit from a slightly more detailed discussion.

As stated before, this finding is in contrast to patients with GLD ( Brown et al, Endocrine 2018) maybe because “patients with PLD are older at the time of metreleptin treatment, thus insulin use may be an indicator of beta-cell failure in PL, whereas it may be an indicator of more severe insulin resistance in GL.”(Adamski, Cur Med Res Opin 2021) This extra explanation might be helpful for understanding the text better.

• Maybe one or two more comments on Supplemental Table 4 (thank you for providing this special information!) are worthwhile. Given the extreme high costs of metreleptin and the burden this provides for the health system, it is an important finding that even in some patients with GLD the outcome does not differ between metreleptin treatment and GLP1 treatment.

I would rather argue the other way around than the authors – it is not surprising that we see no improvement under GLP1 compared to metreleptin (which was stopped -not given additionally)– much more that we see no worsening of the metabolic situation. Admittedly there is no difference to baseline for either treatment, so maybe these patients do not benefit from any intervention.

This could be a problem of numbers or of compliance. The data might also be difficult to interpret because patients with LD and PLD are included.

Thank you for providing these valuable improvements and clarifications.
This article will provide important information to all physicians treating patients with lipodystrophy and is a great achievement.