Journal list menu

Free Access

Treatment-induced neuropathy of diabetes: an update

Juan Eduardo Quiroz-Aldave MD

Juan Eduardo Quiroz-Aldave MD

Division of Medicine, Hospital de Apoyo Chepén, Chepén, Perú

Search for more papers by this author
Marcio José Concepción-Zavaleta MD

Corresponding Author

Marcio José Concepción-Zavaleta MD

Carrera de Medicina Humana, Universidad Científica del Sur, Lima, Perú

Correspondence to: Dr Marcio José Concepción-Zavaleta, 2150 Simon Bolivar Avenue, 15175, Lima, Perú; email: [email protected]Search for more papers by this author
María del Carmen Durand-Vásquez MD

María del Carmen Durand-Vásquez MD

Division of Family Medicine, Hospital de Apoyo Chepén, Chepén, Perú

Search for more papers by this author
Elman Rolando Gamarra-Osorio MD

Elman Rolando Gamarra-Osorio MD

Division of Endocrinology, Hospital Víctor Lazarte Echegaray, Trujillo, Perú

Search for more papers by this author
Rosa María Alcalá-Mendoza MD

Rosa María Alcalá-Mendoza MD

Division of Physical Medicine and Rehabilitation, Hospital Víctor Lazarte Echegaray, Trujillo, Perú

Search for more papers by this author
Susan Luciana Puelles-León MD

Susan Luciana Puelles-León MD

Division of Internal Medicine, Hospital de Apoyo Chepén, Chepén, Perú

Search for more papers by this author
Sofía Ildefonso-Najarro MD

Sofía Ildefonso-Najarro MD

Division of Endocrinology. Hospital Nacional Guillermo Almenara Irigoyén, Lima, Perú

Search for more papers by this author
Luis Alberto Concepción-Urteaga PhD

Luis Alberto Concepción-Urteaga PhD

School of Medicine, National University of Trujillo, Trujillo, Perú

Search for more papers by this author
Jorge Estanislao Gonzáles-Mendoza MD

Jorge Estanislao Gonzáles-Mendoza MD

Alexander von Humboldt Institute of Tropical Medicine, Peruvian University Cayetano Heredia, Lima, Perú

Search for more papers by this author
José Paz-Ibarra MSc

José Paz-Ibarra MSc

School of Medicine, National University of San Marcos, Lima, Perú

Search for more papers by this author
First published: 06 December 2023


Background and aims: Treatment-induced neuropathy of diabetes is an acute small-fibre neuropathy associated with rapid glycaemia improvement.

Methods: This study is a narrative review carried out based on a bibliographic review, using articles indexed in PubMed/Medline and Scielo.

Results: This entity is more frequent in adult patients with poor previous glycaemic control. Its precise pathophysiology is unknown, but it is likely related to unrestored microcirculation changes that occurred during the hyperglycaemic period. It presents with intense, sudden neuropathic pain and autonomic dysfunction after a rapid glycaemic correction and a poorer analgesic response than in diabetic neuropathy.

Conclusions: Since rapid glycaemia correction is the cause of this problem, clinical practice guidelines that can help physicians to prevent, diagnose and manage this entity should be developed. Copyright © 2023 John Wiley & Sons.


Diabetic neuropathy (DN) is one of the main complications of diabetes mellitus (DM), and it is related to poor glycaemic control.1 Its prevalence in persons with DM is approximately 50%, being similar for type 1 and type 2 DM.2 The most common form is distal symmetric polyneuropathy, which accounts for almost 75% of all cases; however, there are some less frequent forms.3-5 One of them is the treatment-induced neuropathy of diabetes (TIND), which is an acute and painful small-fibre neuropathy associated with rapid glycaemic control, and mainly observed in patients with poor previous glycaemic control.6 This condition was first described in 1933, in a female with diabetes who developed severe burning pain after starting insulin therapy. The condition remitted after suspending insulin, and it reappeared after restarting insulin. Back then, it was thought that it was an allergic reaction to insulin, and it was named ‘insulin neuritis’.7 Many similar cases have been reported since then, with the same naming, and others have been classified as ‘acute painful neuropathy’ or ‘diabetic neuropathic cachexia’.8, 9

It was in 2010 that ‘treatment- induced neuropathy of diabetes’ was suggested as a more adequate denomination for this condition since it affects DM patients treated with insulin and oral hypoglycaemic agents and even with a severely restrictive diet; moreover, some of these affected patients developed not only painful peripheral neuropathy but also autonomic neuropathy.10

The objective of this updated review is to describe the epidemiology, pathophysiology, diagnosis, therapy, and perspectives of TIND, aiming to remind our colleagues about this condition and limit the quality-of-life deterioration it brings.


Search strategy

A narrative review was performed, so we started searching papers published in PubMed/Medline and Scielo databases for the period between 2005 and 2022. Papers should be written in Spanish or English. MeSH (Medical Subject Heading) term used was ‘diabetic neuropathies’. We also used ‘insulin neuritis’ as a search term because this is not included as a MeSH term. The search was restricted to studies undertaken in humans.


A total of 12,887 citations were identified. Most of them discussed diabetic neuropathy broadly but did not refer specifically to TIND. Sixty-five papers complied with the inclusion criteria. This process is shown in Figure 1.

Details are in the caption following the image
Flowchart of the narrative review process

Data extraction and synthesis

The results of included papers were carefully assessed including topics regarding epidemiology, pathophysiology, diagnosis, and therapy.

Quality assessment

The quality of our narrative review was assessed using the SANRA scale, which comprises the following items: an explanation of the importance of this review and a description of the literature search, references, scientific reasoning, and appropriate data presentation. The score was 12 points, corresponding to all the required items.11


During the years after its first description, few TIND cases were reported; however, its presence in the medical literature has increased in the last decades12 (Appendix 2), indicating that this condition would be more frequent than it was initially thought.13 Its prevalence in subjects with diabetes in a third-level reference centre was reported to be 10.9%.13

Cases reported in the literature4, 14-31 show that the mean age of those affected by TIND is 36.55 years (95% CI 30.99–42.11), and that there is a slight male predominance (57.14% vs 42.86%). Type 1 DM was more frequent than type 2 DM (64.29% vs 35.71%), and patients had the disease for a mean of 4.9 years (95% CI 3.17–6.64). Four-fifths of all patients (81%) used insulin for treating their DM and 14.3% used oral hypoglycaemic agents only. One case associated with pancreatic islet transplantation was reported. In patients with TIND, HbA1c reduction was on average 7.22% (95% CI 6.11–8.33%) [55mmol/mol (95% CI 43–68)] in a 5.4-week period (95% CI 4.4–6.4 weeks), during which time painful neuropathic symptoms and autonomic manifestations occurred. Some patients used a single drug for alleviating their symptoms, and others used drug combinations: 50% of patients used anticonvulsant drugs (gabapentin and pregabalin) for their therapy, almost 40% used antidepressant agents, and nearly 25% used analgesics. Total or nearly total remission was achieved in 60% of all the cases, 38% had partial remission, and one patient did not achieve symptom remission.

Considering the frequency of type 1 and 2 DM in patients with TIND, and that the prevalence of type 2 DM is approximately 15 times that of type 1 DM,32 we may estimate the risk for the occurrence of TIND to be nearly 30 times higher in individuals with type 1 DM than in those with type 2 DM.


Until recently it was thought that TIND was an allergic reaction to insulin.7 TIND pathophysiology has not been completely elucidated, but some hypotheses are trying to explain this issue.5, 33, 34 One of these is related to changes in microcirculation that occur during hyperglycaemia periods and that do not restore as fast as glycaemia level corrections. This proposed mechanism involves the formation of arteriovenous shunts that may lead to relative hypoxia or ischaemia within the endoneurium.26, 33-35 Changes in the retina and the kidneys that may simultaneously occur support this microvascular mechanism.26 Other hypotheses include cell apoptosis as a result of relative glucose deprivation,18, 29, 36 and areas with axonal regeneration that are formed once normoglycaemia is reached, and these may lead to ectopic depolarisation and pain.16, 35 One less robust hypothesis contemplates an immune reaction against insulin, considering that type 1 DM is likely to be an autoimmune condition and that the frequency of TIND is higher in this form of DM.13, 19 Figure 2 summarises the hypotheses regarding the origin of TIND.

Details are in the caption following the image
Proposed pathophysiological mechanisms for treatment-induced neuropathy of diabetes

TIND must be differentiated from hypoglycaemic neuropathy, another form of neuropathy described in patients with diabetes who use hypoglycaemic agents, in which hypoglycaemic state leads to distal axonopathy including both degenerative and regenerative events, the motor axons being more vulnerable than sensory ones. The cellular mechanisms behind the development of hypoglycaemic disorders of the peripheral nervous system are unknown, hence more studies are needed to elucidate the pathophysiology of hypoglycaemic neuropathy.37


Clinical manifestations of TIND include burning and stinging neuropathic pain, depending on length, that suddenly occurs two to six weeks after the improvement of glycaemic control. Should there be a large reduction of mean glucose levels, proximal and distal neuropathic pain may ensue; if the reduction is not that marked, neuropathic pain may only be distally located.38 Most patients present with autonomic dysfunction (orthostatic hypotension, syncope, early satiety, excess or reduced sweating, and sexual dysfunction) that may occur simultaneously or shortly thereafter, but this may be unnoticed because of the severity of neuropathic pain.13, 21, 38 The severity of autonomic dysfunction is also correlated with the extent of the reduction of mean glucose levels.13

Neurologic examination usually is normal, except for allodynia and hyperalgesia. Nerve conduction studies are usually normal. Some elderly patients with diabetes may present with mild features of pre- existing neuropathy, with absent patellar reflex, and reduced distal vibrating sensitivity.10

Besides neuropathy, many patients with TIND may show rapid progression of other microvascular complications (retinopathy and nephropathy), with the occurrence of neuropathic pain.10, 39 There may likely be a common pathophysiologic pathway involving microcirculation.

In 2015, three criteria for diagnosing TIND were proposed:13
  • ≥2% HbA1c reduction over three months (or a proportional equivalent in time).
  • Occurrence of neuropathic pain and/or autonomic dysfunction within eight weeks after HbA1c reduction.
  • Acute onset of neuropathic pain (more than a three-point increase in the 11-point Likert scale) and/or autonomic dysfunction lasting for more than two weeks, and severely enough for seeking medical care.

Two other important characteristics associated with TIND are the presence of very high HbA1c levels (>10% [86mmol/mol]) before starting therapy and very fast glycaemic control (in a few weeks) after having started therapy.13 A reduction greater than 2% compared to baseline HbA1c values is a strong functional predictor for TIND.40

At the beginning in its natural history, TIND may occur as an acute small-fibre sensitive neuropathy;41 therefore, the main differential diagnoses include variants of Guillain–Barré syndrome,42 steroid-susceptible small-fibre sensorial neuropathy,43 and non-systemic vasculitic neuropathy.44

Electroneuromyography can be used as a complementary tool in the diagnostic evaluation of patients with suspected TIND; sometimes being able to find evidence of sensorimotor polyneuropathy, but at other times obtaining normal results which do not exclude the diagnosis of TIND, because this technique is not totally appropriate to establish a diagnosis of small-fibre neuropathy.45 Skin biopsy with determination of intraepidermic nervous fibre density remains the gold standard method to diagnose small-fibre neuropathy.45 Moreover, corneal confocal microscopy has been proposed as an alternative, non-invasive method; however, its use is not widely available and is restricted to specialised centres.46


A few studies showed that reduction in the insulin doses allowing a permissive hyperglycaemic metabolic state could reduce pain;13 however, the long-term complications remain because of the chronic hyperglycaemia.47

Since the main reason for visiting a physician is neuropathic pain, the main objective of the therapy is pain management; therefore, the same drugs prescribed for DN may be used.33 Although the evidence does not point toward relaxing glycaemic control aiming to accelerate pain resolution, the current consensus states that good glycaemic control must not be relaxed.12, 16, 42, 43

Neuropathic pain in patients with TIND may progressively alleviate in a few weeks or in up to 36 months;30, 41 however, it is necessary to start analgesic therapy to reduce pain.10, 48-50

In some cases, TIND pain is much more severe than that of DN, so single- agent therapy may be maximised, or a second or a third drug may be added to analgesic therapy.47, 51

A poor response to opioids has been observed;13 however, should pain persist, these drugs could be used additionally to other compounds.29

Tricyclic antidepressants may be included in drug therapy, but these should be used with caution, since their anticholinergic adverse effects may worsen TIND-associated orthostatic hypotension.33

Table 1 shows the drugs used in TIND.6, 20, 24, 32, 48-50, 52 Use cautiously in kidney disease and adjust for renal function as needed.

Table 1. Drugs used for managing pain in treatment-induced neuropathy of diabetes, and general dosing
Drug Class Objective Dose (mg/day) Effect (weeks)
Pregabalin First-line anticonvulsant α2-δ ligand 300–600 8–12
Gabapentin Second-line anticonvulsant GABA 900–3600 4–6
Duloxetine First-line antidepressant SNRI 40–60 10–12
Amitriptyline Second-line antidepressant SNRI 25–100 2–4
  • GABA = gamma-aminobutyric acid; SNRI = serotonin and norepinephrine reuptake inhibitors.


The ACCORD trial showed that intensive management of DM, aiming to reach HbA1c normal level, increased mortality and did not significantly reduce major cardiovascular events;53 therefore, the benefit of aggressive therapy for DM may be smaller than its associated risks, and TIND is an excellent reminder of this situation.4 Not all persons with DM get benefit from rapid glycaemia correction, so clinical practice guidelines should pay greater attention to this condition, establishing therapy goals limiting an excessive HbA1c reduction in a determined time period (2% as maximum in three months) and then allowing a safer approach.54 More studies are needed to identify the most adequate way for achieving glycaemic control without increasing the risk of the occurrence of TIND and worsening retinopathy and nephropathy.55, 56

There is still some doubt regarding this issue, such as how long should hyperglycaemia last for reaching the risk of occurrence of TIND and whether there may be other comorbidities that may increase or reduce the risk of developing TIND.

Another point to be considered is whether a vegetarian diet may be a therapeutic option for those who are at risk for developing TIND since the improvement in glycaemia is gradual and not abrupt.57


All authors recognise that despite describing the methodology used for the development of this narrative review, a research question is not answered; therefore, the scientific evidence is less than a systematic review. Additionally, minor limitations include the absence of clinical practice guidelines related to the diagnosis and management of TIND and the lack of epidemiological studies that describe its current prevalence worldwide.


TIND is a complication of intensive control of diabetes, in which, despite its prevalence, there is still not enough scientific evidence for guiding its management; therefore, it is necessary to look for the most adequate way of reaching a good glycaemic control without increasing the risk for the occurrence of TIND and, anticipating this, preventing poor diabetes control, which is the primary cause of TIND.

Declaration of interests

There are no conflicts of interest declared.


  • Treatment-induced neuropathy of diabetes (TIND) is associated with rapid glycaemic improvement in patients with poor previous glycaemic control
  • TIND pathophysiology is not completely clear, but may be related to changes in microcirculation that occur during hyperglycaemia periods
  • TIND diagnosis is clinical but skin biopsy remains the gold standard method to diagnose small-fibre neuropathy
  • Treatment of TIND includes the use of the same drugs prescribed for painful diabetic neuropathy and the maintenance of adequate glycaemic control

Appendix 2: Case reports of treatment-induced neuropathy of diabetes.4, 14-31

Case Age (years) Sex Years of DM Type of DM DM treatment HbA1c reduction % (mmol/mol) Time (weeks) to develop neuropathy Reversion Neuropathy treatment Authors
1 54 F 2.00 2 Insulin 9.60 (81) 8 Partial Acetaminophen, imipramine Tesfaye S, et al. (1996)26
2 52 M 5.00 2 Insulin 3.90 (19) 6 Complete Acetaminophen, codeine Tesfaye S, et al. (1996)26
3 62 M 5.00 2 Insulin 8.50 (69) 4 Partial Imipramine Tesfaye S, et al. (1996)26
4 34 M 0.17 1 Insulin Not specified 4 Complete Acetaminophen, codeine Tesfaye S, et al. (1996)26
5 35 F 0.17 1 Insulin 9.60 (81) 8 Complete Not specified Tesfaye S, et al. (1996)26
6 14 M 1.08 1 Insulin 6.50 (48) 6 Partial Gabapentin, acetaminophen Wilson JL, et al. (2003)19
7 30 M 15.00 1 Insulin 2.20 (1) 3 Complete Amitriptyline Song KB, et al. (2009)28
8 29 M 3.00 1 Insulin 13.90 (128) 4 Complete Carbamazepin, tramadol Dabby R, et al. (2009)18
9 30 F 4.00 1 Insulin 11.80 (105) 2 Complete Gabapentin, duloxetine Dabby R, et al. (2009)18
10 27 M 4.00 1 Insulin 6.20 (44) 3 Complete Duloxetine, pregabalin Dabby R, et al. (2009)18
11 56 M 5.00 2 OADs 4.60 (27) 2 Complete Amitriptyline, gabapentin, pregabalin, duloxetine Dabby R, et al. (2009)18
12 58 M 5.00 2 Glibenclamide 6.70 (50) 4 Complete Carbamazepine, amitriptyline, duloxetine Dabby R, et al. (2009)18
13 52 M 5.00 2 OADs 9.10 (76) 2 Complete Duloxetine, gabapentine, oxycodone Dabby R, et al. (2009)18
14 22 F Many years (not specified) 1 Insulin Not specified 4 Complete Gabapentin, opioids Smith AG, et al. (2012)25
15 23 F Not specified 1 Insulin Not specified 5 Partial Not specified Gibbons CH. (2014)21
16 57 M 5.00 2 Not specified 4.00 (20) 4 Complete Diet Tran C, et al. (2015)4
17 44 M 0.50 1 Insulin 7.90 (63) 6 Partial Pregabalin, gabapentin, duloxetine, opioids Tran C, et al. (2015)4
18 49 M 0.33 2 Insulin 9.00 (75) Not specified Complete Duloxetine Tran C, et al. (2015)4
19 41 F 0.08 1 Insulin 9.10 (76) 4 Partial Pregabalin, amitriptyline Tran C, et al. (2015)4
20 22 F 16.00 1 Insulin 1.10 (-) 16 Partial Gabapentin Hwang YT, et al. (2016)16
21 24 F 16.00 1 Insulin 5.20 (33) 2 Complete Pregabalin Hwang YT, et al. (2016)16
22 43 F 16.00 1 Insulin 8.40 (68) 4 Complete Pregabalin Aladdin Y, et al. (2017)20
23 46 M 14.00 2 Metformin, glibenclamide 12.10 (109) 4 Partial Gabapentin Aladdin Y, et al. (2017)20
24 54 F 2.00 2 Insulin

2.70 (6)


Complete Human immunoglobulin Duarte JM, et al. (2018)15
25 25 M 0.25 1 Insulin 6.20 (44) 6 Partial NSAIDs, tramadol, duloxetine Cuenca Hernández R, et al. (2018)14
26 35 F 23.00 1 Pancreatic islet transplant 9.10 (76) 1.5 Non-remission Pregabalin, duloxetine, venlafaxine, amitriptyline Meillet L, et al. (2019)27
27 16 F 1.50 1 Insulin 10.40 (90) 4 Complete Amitriptyline, pregabalin and gabapentin Varadharaju N, et al. (2019)29
28 56 M 0.67 2 Insulin 8.10 (65) 6 Complete Pregabalin Siddique N, et al. (2020)30
29 67 M 1.00 2 Insulin, metformin, glicazide 9.40 (79) 8 Partial Pregabalin Siddique N, et al. (2020)30
30 58 M 2.00 2 Dapagliflozin Not specified 2 Complete None Siddique N, et al. (2020)30
31 19 F 5.00 1 Insulin 8.80 (73) 6 Partial Not specified Yuan J, et al. (2020)17
32 81 F 12.00 2 Metformin Not specified 0.5 Complete Not specified Yuan J, et al. (2020)17
33 18 M 1.00 1 Insulin 7.30 (56) 5 Partial Pregabalin Chandler E, et al. (2020)22
34 14 M 0.25 1 Insulin 8.40 (68) 12 Complete Gabapentin Alexandrou EG, et al. (2020)24
35 16 F 0.08 1 Insulin

4.10 (21)

4 Partial Gabapentin Alexandrou EG, et al. (2020)24


M 5.00 1 Insulin 3.50 (15) 4 Complete Gabapentin Alexandrou EG, et al. (2020)24
37 22 M 3.00 1 Insulin 3.50 (15) 8 Complete Duloxetin Alexandrou EG, et al. (2020)24
38 19 M 3.00 1 Insulin 5.70 (39) 8 Partial Not specified Alexandrou EG, et al. (2020)24
39 9 F 0.17 1 Insulin 8.00 (64) 4 Complete Not specified Alexandrou EG, et al. (2020)24
40 17 M 0.08 1 Insulin 3.30 (13) 4 Complete Gabapentin and cannabis Alexandrou EG, et al. (2020)24
41 20 F New onset 1 Insulin 10.30 (89) 12 Partial Venlafaxine Stainforth-Dubois M, et al. (2021)23
42 66 F 9 2 Insulin, metformin 8.90 (74) 12 Partial Not specified Broadhead DY, et al. (2021)31
  • DM = diabetes mellitus; M = male; F = female; OADs = oral antidiabetic drugs; NSAIDs = non-steroidal anti-inflammatory drugs.