Deciphering Bone Microarchitecture in Diabetic Charcot Neuroarthropathy of Foot: A Case Control Study

Abstract

Introduction:

Charcot neuroarthropathy (CNO) of foot characterised by an increased bone turnover denoted by serological markers of bone resorption. However, histological characteristics of foot bones in people with CNO are not well elucidated.

Methods:

The foot bone samples were collected from patients who had either surgical reconstruction or below-knee amputations for chronic CNO foot (n = 10, Group A), unsalvageable diabetic foot ulcer (n = 16, Group B), and non-diabetic healthy controls following road traffic accident (n = 16, group C). Calcaneum bones retrieved were processed and sections (Haemotoxylin and Eosin, Masson-Goldner stain) evaluated for quantitative histopathological parameters including bony trabeculae number, trabeculae thinning, osteoclast number, Howship’s lacunae, and Haversian canal.

Results:

The mean age of participants in the CNO group was 61.6 ± 5.0 and 62.9 ± 6.5 years in diabetic neuropathy group with duration of diabetes 13.1 ± 6.8 and 14.1 ± 9.1 years with HbA1c of 7.6 ± 1.8% and 8.7 ± 2.6 in group A and B, respectively. We observed that normal bone trabeculae were 15% (10–37.5) in group A and 60% (47.5–82.5) in group B as compared to controls (P = <0.001). Thin bone trabeculae (%) were observed in 10% (3.5–77.5) and 7.5% (0–30), P =<0.001), with increased Howship’s lacunae number (1.5 [0.25–2] and 1 [0–2.25] (P = <0.001)) and increased osteoclast number in group A and B as compared to healthy controls.

Conclusions:

There is an increased bone resorption in CNO causing thinning of bone trabeculae secondary to increased osteoclast numbers and Howship’s lacunae in CNO of foot. Anti-resorptive therapies that target osteoclast activity may be an appealing treatment option for diabetic CNO of foot.

INTRODUCTION

Charcot neuroarthropathy (CNO) of foot in diabetes is associated with foot bone fractures, deformities, recurrent ulcers, limb amputation, and higher mortality.[1,2] CNO of foot usually presents with localised swelling, warmth, and erythema of the foot.[3] The pathogenesis of CNO involves peripheral sensory neuropathy leading to loss of protective sensation, susceptibility to repetitive trauma, and abnormal bone metabolism, ultimately causing joint disintegration and foot deformity.[4,5] The conceptual understanding of the pathogenesis of CNO has evolved over time from the neurotraumatic and neurovascular theory to the recognition of osteoclastic resorption of foot bones following activation of the receptor activator of nuclear factor kappa-B (RANK).[6,7] The CNO is associated with an increase in markers of bone turnover (BTM), including formation (P1NP, ALP) and resorption markers (CTx, de-oxy pryridonilines), as well as inflammatory cytokines (IL-6, TNF-a).[8,9] Periodic analysis of BTMs in active CNO suggests osteolysis of bones consistent with acute resorption of bones followed by sclerosis, particularly along joints following remission of CNO. However, the microstructural properties of foot bones following the CNO process are unknown.

The quality of foot bones, especially the histological characteristics that could help in better understanding the pathophysiology of CNO, is not explored. A study conducted in animal models revealed an increase in disorganized bone and early subchondral hypertrophy of woven bone.[10] La Fontaine et al. found that foot bones in CNO were infiltrated with inflammatory myxoid tissue and had a disorganised trabecular pattern.[11] Dharmadas et al. described a decreased number of osteocytes and plenty of empty lacunae in a study of six patients of CNO compared with two non-diabetic controls.[12] Recently, the FEMASK score developed by King et al. on histomorphometric analysis of mid-foot bones has provided insights into the role of intraneural vasculopathy in destructive CNO. However, it did not address microarchitectural changes including osteoclasts and bone resorption details, especially in calcaneus, which is the critical load-bearing bone frequently affected in chronic, non-salvageable Charcot foot.[13]

The aim of the present study was to conduct a quantitative analysis of the histological parameters of foot bone in patients with CNO compared with diabetic neuropathy and non-diabetic healthy controls.

MATERIALS AND METHODS

Study participants

All patients presenting to the diabetic foot clinic with either CNO of the feet or neuropathic diabetic foot ulcer (DFU) were screened. Charcot foot ([Hind Foot]) and neuropathic DFUs that underwent foot reconstruction surgery or were offered below knee amputation (BKA) in view of unsalvageable limbs decided by a multidisciplinary team including orthopaedicians, surgeons, and endocrinologists were included in the study. Participants without pre-existing diabetes undergoing BKA due to road traffic accidents were taken as healthy controls after ruling out prior diabetes or prolonged steroid usage or on any medications affecting bone turnover.

We recruited patients from February 2022 to August 2024. Experimental procedures and data analysis were performed over the course of one and half year following recruitment. During the study period, 56 patients were identified and 42 were included in the analysis (Charcot foot [n = 10, group A], neuropathic DFU [n = 16, group B], and non-diabetic controls [n = 16, group C]) explained in Figure 1. In light of the study’s exploratory design, convenience sampling was chosen as a pragmatic approach to recruit participants efficiently. All patients were euthyroid, and groups were matched in terms of hypogonadism or premature menopause. Patients with osteoporosis (T score -2.5 at the lumbar spine or hip) who had received anti-resorptive agents or anabolic therapies in the previous 12 months, steroid in the past 3 months, estimated glomerular filtration rate <30 ml/min/m², significant peripheral vascular disease (ankle-brachial index <0.6), and primary or secondary hyperparathyroidism, people with diabetes on glitazones and pregnant, and lactating women were excluded from the study.

Figure 1.

Characteristics of the study population in three groups

Sample procurement and storage

Calcnaeum bone is predominantly trabecular bone (80%) and is a weight bearing bone that provides vital insights across bone mineral and metabolic disorders including in orthopaedics, radiology, anthropology, and sports medicine. Therefore, calcaneum bone was collected in present exploratory study from patients undergoing surgical procedures for foot complications. A 1–3 cylindrical trabecular bone cores were extracted from calcaneum, each of 10–12 mm in diameter and 0.5–1 cm in length by using a Gotheur needle (12 G). The core of the retrieved bones was cleaned with normal saline, cut into smaller bone chips, fixed into two parts, and processed for further histopathological procedures by Haematoxylin and Eosin and Goldner stain. The histopathological evaluation involved two independent investigators.

Haematoxylin and eosin stain procedure

Small bone chips were initially fixed in a 10% neutral buffered formalin solution for 48 h to prevent autolysis and maintain cellular integrity. Subsequently, decalcification was performed to remove mineral content from the bones. Dehydration of the specimens was then carried out using a sequential immersion (Sakura Tissue-Tek Rotary Tissue Processor. 4640, USA) in 50%, 70%, and 90% ethanol solutions to eliminate water and unbound fixatives. Following dehydration, cleansing steps were conducted utilising Xylene 50%, 70%, and 90% as clearing agents to displace residual dehydrating solutions and enhance tissue receptivity to the infiltrating medium. Subsequent to clearing, the bone tissue sections were impregnated with paraffin wax to provide structural support, enabling the production of thin sections for analysis. The paraffin-embedded (Tissue-Tek TEC™ 6 Embedding Console System Sakura, USA) bone was then sectioned into 4 micrometers slices using a Rotary microtome (Microm Rotary Microtome, HM355, USA). Every eighth section, resulting in a total of four sections per slide, was selected and isolated for subsequent analysis.

Following isolation, the slides underwent a dewaxing process. Initially, immersion in xylene for 5 min was followed by sequential exposure to 90% and 70% absolute alcohol for 10 s each. Subsequently, the slides were rinsed with water. Haematoxylin staining ensued for 8 min, followed by washing with distilled water. Stain discolouration via 1% acetic acid alcohol (differentiation) was carried out, followed by further washing with distilled water. Eosin staining was then performed for 5 s, followed by water rinsing. Post staining, slides were dehydrated with alcohol (70%, 90%, 100%) for 10 s each, followed by drying in a hot air oven for 2–3 min. Subsequently, slides were cleared with xylene for 2–3 min. Finally, mounting of the slides was conducted for observation under a light microscope (Olympus Trinocular Research Microscope Model BX 51 with Auto-system PM10SP).

Masson’s Goldner stain procedure

The bone chips were initially fixed in 75% absolute alcohol for 24 h, followed by immersion in a 2% silver nitrate (AgNO3) solution in darkness for 48 h. Subsequently, the samples underwent thorough washing under running tap water. Next, the bone chips were immersed in aqueous sodium hypophosphate (0.1N NaOH) for 48 h, followed by transfer to a 5% aqueous sodium thiosulfate (Na2S2O3) solution for an additional 24 h. After further washing, the samples underwent routine decalcification using formic acid for 2 days.[14] The subsequent embedding, cutting, and dewaxing procedures were conducted in accordance with previously described methods.

The initial step involved the application of Weigert’s iron haematoxylin solution to the slides for a duration of 5 min, followed by rinsing with running tap water. Subsequently, they were subjected to a 1% acetic acid solution wash for 30 s. Following this, staining with Azophloxine solution was performed for 12 min, followed by another wash with a 1% acetic acid solution. The slides were then stained with Tungtophosphoric acid orange G solution for 1 min and underwent an additional wash with a 1% acetic acid solution. Next, staining with light-green SF solution was executed, succeeded by clearing with 100% ethanol for 2 min. The finalised slides were mounted and observed under a light microscope (Olympus microscope ×10 and ×20).

Each case involved eight slides in total—four stained with H and E and four stained with the Masson-Goldner stain—and histopathological parameters were calculated by taking the mean of four slides for each parameter of the respective stain. Every specimen underwent evaluation by a pathologist who was not aware of the groups. We studied the following histopathology parameters: Bone trabeculae (BT) classified as Normal (N), Thin (T) as compared to control, Bony trabeculae fragmentation (F%), Osteoblast rimming (OBR) present or absent, osteoclast reaction present or absent, osteoclast number (OC), Howship’s lacunae number (HL), and Haversian canal number (HC), which were calculated by taking average of observed four slides per field.

Statistical analysis

Statistical analysis was conducted with IBM SPSS version 29 (IBM Corp., Armok, NY, USA). The normality of the data was determined by Kolmogorov–Smirnov test. Inter-group comparisons were done with Kruskal–Wallis test with bon-Ferroni correction for variables not normally distributed amongst the three groups. Mann–Whitney test was performed for pairwise comparison. A P < 0.05 was considered significant.

Ethical aspects

Ethical Committee name: Postgraduate Institute of Medical Education and Research, Chandigarh Institutional Ethics Committee, Institute: Postgraduate Institute of medical education and research, Ethical Clearance no: PGI/IEC/2022/EIC000232 dated 18.02.2022. Informed consent was taken from all the participants, and they were informed that the information would be used for research and educational purpose. All the procedures followed the guidelines laid down in Declaration of Helsinki 1964.

RESULTS

The demographic and laboratory parameters are provided in Table 1. A substantial proportion of patients (Group A and B) presented with comorbidities, including hypertension (61.5%), obesity (46.2%), chronic kidney disease (26.9%), retinopathy (23.1%), and coronary artery disease (19.2%).

Table 1.

Characteristics of the study population in three groups

Parameters	Charcot (n=10)	DFU (n=16)	Controls (n=16)	P
Age (years)	61.6±5.05	62.86±6.50	44.12±17.27	<0.001*
Gender
Male	7 (70%)	13 (81.25%)	14 (87.5%)
Female	3 (18.7%)	3 (18.7%)	2 (12.5%)	0.54
Duration of diabetes (years)	13.14±6.81	14.07±9.14	-	<0.001*
HbA1C (%)	7.57±1.79	8.70±2.65	5.35±0.55	<0.001*
Urea (mg/dL)	45.21±18.34	64.45±49.01	28.60±12.65	0.02*
Creatinine (mg/dL)	2.15±1.95	1.95±1.0	0.81±0.24	0.21
eGFR (ml/min/1.73m2)	99.33±143.82	75.16±45.22	110.70±36.76	0.45
VPT SCORE (mv)	50.00±0.00	48.00±5.34	0	<0.001*
ABI (LT)	0.94±0.19	0.7±0.49	0.9	<0.001*
ABI (RT)	1.04±0.13	0.77±0.45	0.9	<0.001*
FBG (mg/dL)	152.91±86.23	178.42±84.36	78.87±23.68	<0.001*
PPG (mg/dL)	223.37±55.51	187.35±97.72	89.0±13.23	<0.001*
Hb (gm/dL)	10.01±1.87	9.14±1.40	10.3±1.6	0.33
LDL (mg/dL)	61.84±42.45	58.47±31.30	78.19±31.39	0.31
HDL (mg/dL)	35.36±8.32	31.03±8.88	52.12±40.60	0.13
TG (mg/dL)	148.04±80.74	111.12±27.90	125.80±80.54	0.52
ALP (U/L)	88.37±70.16	139.46±115.57	94.69±33.55	0.32
Albumin (g/dL)	4.14±2.93	3.92±3.0	3.32±1.55	0.82
Calcium (mg/dL)	8.86±0.89	8.40±0.84	7.59±2.07	0.18
Inorganic phosphorus (mg/dL)	4.76±1.57	3.19±0.85	3.70±1.05	0.02*
P1NP (ng/ml)	146.70±221.04	115.8±133.27	71.20±79.74	0.50
CTX (pg/ml)	797.0±607.70	1026.3±844.33	821.15±653.57	0.73
25 (OH) D D (ng/ml)	26.8±5.50	22.12±18.26	14.89±12.31	0.24
iPTH (pg/ml)	34.88±28.62	32.74±17.12	33.80±15.10	0.97

*P>0.05 and considered significant (P value reflects the 2 df Kruskall–Wallis test or Chi-square test, as appropriate). HbA1c=Haemoglobin A1C, eGFR=Estimated glomerular filtration rate, VPT=Vibration perception threshold, ABI=Ankle-brachial index, FBG=Fasting blood glucose, PPBG=Glucose postprandial blood, Hb=Haemoglobin, CHOL=Cholesterol, LDL=Low density lipoprotein, TG=Triglyceride, HDL=High-density lipoprotein, AST=Aspartate transferase, ALT=Alanine transaminase, ALP=Alkaline phosphate, PINP=Procollagen 1 intact N- terminal propeptide, CTX=C-terminal telopeptide of Type 1 collagen, IPTH=Intact parathormone, 25(OH) D3=Cholecalciferol, DFU=Diabetic foot ulcer

Quantitative HPE findings

We found that the BT (N) was 15% (10–40) in group A compared to 60% (45–85) and 100% (97.5–100), P =<0.001, in group B and C, respectively. The BT (T) was 10% (3–80), (7.5% [0–30], 0% [0–0], P = <0.001), OC (n) (0 [0–1], 0 [0–0], 0 [0–0], P = 0.03), and HL (n) (1.5 [0–2],1 [0–2.5], 0 [0–0]; P = <0.001) in group A, B, and C, respectively.

Table 2 shows that there was no difference in percent bone fragmentation (P = 0.39) or HC number (P = 0.37) in the CNO group as compared to those with DFU or healthy controls. The Masson-Goldner stain illustrated similar findings as observed with H and E stains.

Table 2.

Comparative histopathological analysis of foot bone in different groups using Haematoxylin and Eosin and Masson-Goldner staining

Parameters	Charcot (n=10)	DFU (n=16)	Controls (n=16)	Comparison of Charcot, DFU, and Control Groups (P)	Charcot versus DFU (P)	Charcot versus controls (P)	DFU versus controls (P)
Haematoxylin and Eosin stain (H and E stain)
BT (N) %#	15 (10–40)	60 (45–85)	100 (97.5–100)	<0.001$	0.07	<0.001*	<0.001*
BT (T)%#	10 (3–80)	7.5 (0–30)	0 (0–0)	<0.001$	0.14	<0.001*	<0.001*
BT (F)%#	15 (5–80)	17.5 (10–35)	10 (5–30)	0.39	0.97	0.45	0.15
OBR (n)#	2.5 (0–5)	0 (0–0)	0 (0–0)	<0.001	0.01*	0.01*	0.96
OC (n)#	0 (0–1)	0 (0–0)	0 (0–0)	0.03$	0.10	0.02*	0.31
HL (n)#	1.5 (0–2)	1 (0–2.5)	0 (0–0)	<0.001$	0.58	<0.001*	<0.001*
HC (n)#	2 (2–3)	3 (2–4)	2.5 (1–6.5)	0.37	0.08	0.66	0.63
Masson-Goldner stain
BT (N) %#	5 (0–25)	60 (47.5–80)	100 (100–100)	<0.001$	0.01*	<0.001*	<0.001*
BT (T)%#	10 (0–57.5)	10 (0–30)	0 (0–0)	<0.001$	0.87	<0.001*	<0.001*
BT (F)%#	30 (15–52.5)	20 (10–42.5)	15 (5–47.5)	0.60	0.47	0.36	0.60
OBR (n)#	0 (0–0)	0 (0–0)	0 (0–0)	0.50	0.54	1.00	0.31
OC (n)#	0 (0–07.5)	0 (0–0)	0 (0–0)	0.01$	0.05*	0.01*	0.31
HL (n)#	2 (1.5–2.75)	1.5 (0–3)	0 (0–0)	<0.001$	0.59	<0.001*	<0.001*
HC (n)#	1.5 (1–3)	3 (2–4)	2.5 (1–4.75)	0.63	0.28	0.74	0.61

^$P>0.05 and considered significant using Kruskal–Wallis test, *P>0.05 and considered significant using Mann–Whitney test, ^#Average number of respective parameter per ×10 field. BT=Bone trabeculae, N=Normal, T=Thin, F=Fragmentated, OBR=Osteoblast rimming, OCR=Osteoclast reactions, n=number, HL=Howship’s lacunae, HC=Haversian canal, DFU=Diabetic foot ulcer

Descriptive HPE findings

The moth-eaten appearance of bone was observed in the CNO group unlike the hourglass appearance representing normal bony trabeculae in the control group as shown in Figure 2 (1) and (3). Fragmented bony trabeculae with tunnel effect, dead bone fragments with compensated hypertrophy, and an increased rate of marrow fibrosis were observed in Charcot bone. The number of Haversian canals was higher, and inter-trabecular spaces showed marrow fibrosis with dilated capillary channels in the DFU group compared to CNO or controls [Figure 2 (2) and (3)]. The adjacent marrow was infiltrated with inflammatory cells that were predominantly lymphoplasmocyte-rich in CNO [Figure 2 (2)].

Figure 2.

(1) a-c (H and E stain) Pictures of Charcot group show diffuse thin fragmented bony trabeculae. Dense lymphomononuclear inflammation in the inter-trabecular spaces along with thinning of bone and Howship lacunae formation (×20). Cluster of osteoclastic cells (black arrow, c) residing at the resorption pit (×20. d-f (Masson Goldner Stain) showing diffuse thin bony trabeculae (×10) and osteoblastic rimming (black arrow, f) and Haversian canal formation (×20 magnification). (2) DFU group: g-i (H and E stain) (×4 and × 10) Thin and thick bone trabeculae, Haversian canal, and occasional Howship lacunae (black arrow, H), fragmentation of bone. j-l (Goldner stain) (×4 and × 10) Thin bone trabeculae, fragmentation of bone. (3) Control group: m-o (H and E stain) (×4 and × 10) normal bone architecture, well-organised lamellar bone, and Haversian canal and p-r (Goldner stain) (×4 and × 10) normal bone trabeculae, fragmentation absent. DFU = Diabetic foot ulcer

DISCUSSION

The results of the present study showed significantly increased osteoclastic numbers and osteoclastic activity associated with the thinning of foot bone trabeculae and a reduced number of normal bony trabeculae in CNO compared to those with diabetic neuropathy or healthy non-diabetic controls.

An increased bone turnover in CNO is evident by increased markers of bone turnover and an increase in activity on Tc99m Sestamibi scan and fluoride uptake on positron emission tomography/computed tomography.[8,15,16,17] However, data were lacking regarding the actual histopathological evidence in CNO. We studied histopathological parameters of calcaneum bone and observed marked thinning of bone trabeculae, increased proportion of fragmented bone trabeculae, and increased number of Howship’s lacunae with enhanced osteoclastic reactions in CNO group compared to healthy controls. The percentage of normal bone was less in CNO as compared to DFU and control. Overall, thinned trabeculae in CNO suggest the role of increased osteoclast number and activity as evidenced by the presence of Howship’s lacunae and not in normal controls.

An increased Haversian canal in CNO suggests an increase in marrow vascularity secondary to diabetic neuropathy and inflammation that is characteristic of CNO. Other interesting findings in CNO group include OBR and empty lacunae with the absence of osteocytes. Osteocytes are mature cells that connect the bone surface, osteoblasts, osteoclasts, and other osteocytes through a canalicular network, which is essential for bone remodelling.[18,19] An absence of osteocytes in CNO suggests a poor reparative capacity of the involved bone, which possibly could be the reason for increased fracture and disorganisation of foot joints.

Recently, King et al. have introduced the FEMASK score to characterise the histopathological findings including histiocytic infiltration, periarticular fibrosis, and granulation in midfoot CNO, as also observed in the present study.[13] While their work emphasised the neurovascular component of disease progression, it did not evaluate bone remodelling dynamics or quantify structural bone changes at the microscopic level. The observed increase in osteoclast numbers, Howship’s lacunae, and trabecular thinning underscores the predominance of osteoclastic activity in the pathogenesis of Charcot foot, supporting the potential utility of anti-resorptive therapeutic strategies in advanced stages of the disease.

The present study though suggests an increased osteoclastic activity, but prior trials with anti-osteoclastic agents including bisphosphonates have not been shown to improve chances of remission or remission time in active CNO.[15,20] It suggests that the pathophysiology of CNO is more complex than local inflammation-activated osteoclasts. This also suggests that the drugs previously used to inhibit osteoclast activity for active CNO are not potent to completely inhibit osteoclast activity. However, recently, anti-RANKL antibody denosumab has been shown to improve outcomes in people with active CNO of foot with chronic kidney disease.[4,21]

The strength of present study is that the bone histology of CNO is compared with appropriate controls, namely, diabetic neuropathic and non-diabetic participants. Only calcaneum bone was studied to avoid structural and compositional heterogeneity in the result. The present analysis is the first study to delineate histopathological quantitative data in the absence of generally accepted criteria for histopathological findings in CNO.

Limitations

We noticed certain shortcomings that may limit the generalisability of the results including the age of control group that was younger compared to the other groups. Secondly, the small sample sizes in the three groups may limit the volume of statistical testing and its interpretation. Although the sample size is small, this study provides valuable insights that can be further explored in larger cohorts and extended studies. The present results are limited to only the calcaneal bone as we included hind foot Charcot, though mid foot bones are more commonly affected in CNO. Urine calcium was not performed to rule out aetiologies like idiopathic hypercalciuria that could alter bone turnover. However, calcium profile was performed to exclude hyperparathyroidism, and therapies that could affect bone turnover like steroids, bisphosphonate, or anti-RANKL therapy.

CONCLUSIONS

The present study identifies a significantly increased osteoclast number and activity in the foot bones of patients with diabetic CNO of foot. Hence, therapeutic agents, especially targeting osteoclasts including the anti-RANKL antibody, denosumab, may be utilised for optimising the scope of surgical and interventional treatment in people with CNO. The poor bone quality assessed on histopathology as compared to the controls provides mechanistic insights and confirms the increased prevalence of fracture and deformity of foot in CNO.

Author contributions

AR conceptualized the study, designed study protocol, provided clinical care to the patients, wrote and edited the manuscript. RS involved in analysis of the data and wrote the initial draft of the manuscript, data collection and analysis. SHK and UNS analysed the histopathological data. UCS, SS, RK provided sample and clinical care. AR is the guarantor of the manuscript.

Conflicts of interest

There are no conflicts of interest.

Use of artificial intelligence

No AI was used for this article.

Data availability statement

On request from the editor.

Funding Statement

We thank SERB-DST, New Delhi for funding the study with research grant number CRG/2021/000680 to Dr Ashu Rastogi.