Table 1.

The role of incretins in the prevention of peripheral neuropathy, a summary of research in humans

Level of evidence	Reference	Key characteristics	Key findings
GLP-1RAs
Cohort studies	Issar et al 2021 [93]	N=90. Exenatide vs DPP-4i vs SGLT2i vs healthy control participants for cross-sectional study. Ten patients taking exenatide followed for a further 3 months in a prospective study and compared with 32 control participants.	Improvements in nerve physiological measures with exenatide use compared with no exenatide use, including improved nerve function, S2 accommodation, superexcitabillity and subexcitabilty. These changes were independent of HbA1c reduction.
	Deng et al 2022 [76]	N=8252. Compared insulin glargine, liraglutide, glimepiride and sitagliptin. Propensity scores used to account for differences in baseline characteristic. These were created from several baseline characteristics (including HbA1c, ethnicity and comorbidity).	No significant difference in neuropathy incidence with liraglutide use compared with other medications.
RCTs	Ponirakis et al 2020 [94]	N=38. Three months of pioglitazone and exenatide vs glargine with aspart insulin.	Vibration perception worsened after 3 months’ usage of pioglitazone and exenatide with no change in neuropathic pain.
	Jaiswal et al 2015 [95]	N=46. Exenatide vs insulin glargine.	No significant difference in incident neuropathy or electrophysiology markers.
	Brock et al 2019 [96]	N=36 (type 1 diabetes) with confirmed polyneuropathy. Randomised to liraglutide or placebo.	No effect of liraglutide on measures of nerve function.
	GRADE Study Research Group et al 2022 [77]	N=5047. Compared insulin glargine, liraglutide, glimepiride and sitagliptin.	No difference in incidence of diabetic polyneuropathy between any of the four medications over a median follow-up of 5 years.
Simulated long-term follow-up of RCT	Sullivan et al 2009 [97]	N=746 + three n=5000 hypothetical cohorts. Liraglutide vs glimepiride.	Predicted neuropathy incidence was lower in the liraglutide group.
DPP-4is
Cross-sectional studies	Issar et al 2021 [93]	N=90. Exenatide vs DPP-4i vs SGLT2i vs healthy control participants.	Physiological markers of nerve function were similar in DPP-4i and SGLT2i groups with more favourable findings in the exenatide group.
Cohort studies	Kolaczynski et al 2016 [71]	N=16,321. Propensity score matching to compare vildagliptin use to sulfonylureas. Propensity scores were derived from the probability of treatment assignment from the following baseline factors: age, sex, line of therapy, HbA1c score, duration of disease (<5 years vs ≥5 years), duration of treatment, previous hypoglycaemic events, co-prescribed medications and number of comorbidities.	Lower neuropathy incidence with vildagliptin OR 0.71 (95% CI 0.60, 0.85).
	Deng et al 2022 [76]	N=8252. Compared insulin glargine, liraglutide, glimepiride and sitagliptin. Propensity scores used to account for differences in baseline characteristics. These were created from several baseline characteristics (including HbA1c, ethnicity and comorbidity).	Sitagliptin appeared to be associated with lower neuropathy incidence compared with glimepiride HR 0.87 (95% CI 0.76, 0.99), although this was not statistically significant after the p value was adjusted for multiple testing (p=0.09).
RCTs	da Silva et al 2015 [98] (abstract only)	N=30. Sitagliptin vs NPH insulin in patients with type 2 diabetes of long duration.	No difference in electrophysiological markers of neuropathy at 1 year.
	TECOS trial, original publication: Green et al 2015 [73]	N= 14,671. Sitagliptin vs placebo.	Incidence of diabetic neuropathy 4.1% in Sitagliptin group, 3.8% in placebo.
	TECOS trial, CKD participants only: Engel et al 2017 [99]	N=3324. Sitagliptin vs placebo. Subgroup analysis of patients with CKD.	No significant difference in neuropathy incidence was observed, risk difference for sitagliptin vs placebo using M-N method = 0.21 (95% CI −1.09, 1.53).
	GRADE Study Research Group et al 2022 [77]	N=5047. Compared insulin glargine, liraglutide, glimepiride and sitagliptin.	No difference in incidence of diabetic polyneuropathy between any of the four medications over a median follow-up of 5 years.
	Gabriel et al 2023 [79]	N=658. Linagliptin vs placebo vs metformin vs linagliptin + metformin.	Linagliptin associated with 19.5% (95% CI 10.1, 29.0) reduction in small fibre peripheral neuropathy compared with placebo.

CKD, chronic kidney disease; SGLT2i, sodium–glucose cotransporter 2 inhibitor