Nociception at the diabetic foot, an uncharted territory

Abstract

The diabetic foot is characterised by painless foot ulceration and/or arthropathy; it is a typical complication of painless diabetic neuropathy. Neuropathy depletes the foot skin of intraepidermal nerve fibre endings of the afferent A-delta and C-fibres, which are mostly nociceptors and excitable by noxious stimuli only. However, some of them are cold or warm receptors whose functions in diabetic neuropathy have frequently been reported. Hence, it is well established by quantitative sensory testing that thermal detection thresholds at the foot skin increase during the course of painless diabetic neuropathy. Pain perception (nociception), by contrast, has rarely been studied. Recent pilot studies of pinprick pain at plantar digital skinfolds showed that the perception threshold was always above the upper limit of measurement of 512 mN (equivalent to 51.2 g) at the diabetic foot. However, deep pressure pain perception threshold at musculus abductor hallucis was beyond 1400 kPa (equivalent to 14 kg; limit of measurement) only in every fifth case. These discrepancies of pain perception between forefoot and hindfoot, and between skin and muscle, demand further study. Measuring nociception at the feet in diabetes opens promising clinical perspectives. A critical nociception threshold may be quantified (probably corresponding to a critical number of intraepidermal nerve fibre endings), beyond which the individual risk of a diabetic foot rises appreciably. Staging of diabetic neuropathy according to nociception thresholds at the feet is highly desirable as guidance to an individualised injury prevention strategy.

Key Words: Foot ulcer, Neuroarthropathy, Insensitivity to pain, Pain perception, Diabetes mellitus, Amputation, Diabetic neuropathy

Core tip: The diabetic foot is characterised by painless ulcers and/or arthropathy. Although painless diabetic neuropathy is known as the underlying condition, little is known quantitatively about the pain evoked by noxious stimuli (nociception) at the diabetic foot. Preliminary evidence shows that pinprick pain perception threshold at plantar digital skinfolds is supranormal, beyond the upper limits of measurement. It is suggested that measuring nociception at the foot in diabetes could specify the individual risk of painless ulcers and/or arthropathy and, thereby, provide the basis of an individualised graded injury prevention strategy.

INTRODUCTION

The diabetic foot (also called diabetic foot syndrome, or diabetic podopathy) is characterised by painless foot ulceration and/or arthropathy in diabetes. Foot injuries and inflammations do not hurt, as illustrated on Figure 1.

Figure 1 Sketch reproduced from Brand P[2].

Painless diabetic neuropathy (PLDN) is the underlying condition. Skin ulcers extending to the subcutaneous tissues are completely anaesthetic, while arthropathy may display faint deep dull aching upon load bearing. Nociception, that is the perception of pain originating from neural processes of encoding and processing noxious stimuli, is failing at the diabetic foot. Although insensitivity to pain at the diabetic foot is common knowledge, details are largely unknown[1,2].

Research into painlessness in diabetes is scarce. Early studies date back to Pamela Margaret Le Quesne[3] and her group almost 30 years ago. They tried to measure the pain perception (nociception) at the diabetic foot by pinching the skin with a custom made “pinchometer”. The results were inconclusive at best[4]. Other authors designed calibrated tools for assessing hypersensitivity of so-called symptomatic, i.e., painful, diabetic neuropathy (SDN, see below). To this end, pinprick pain perception, axon-reflex reaction and temperature detection of the skin were studied. These modalities represent the functions of the afferent A-delta and C-fibres (so-called small fibres; see below) whose contributions to SDN, however, remain controversial[5,6]. SDN will not be discussed here.

To date, measuring pain perception (i.e., nociception) at the diabetic foot was deemed futile with the excuse that “The threshold of sensation that protects normal feet from injury is difficult to define…it is extremely difficult to define a ‘significant loss of sensation’, or at what level sensory loss becomes ‘critical’..”[7]. However, there is some dissent on this issue, as expressed by Dyck et al[8]: “Sadly, the clinical assessment of decreased sensation by physicians is generally inadequate because it is not performed or is performed badly.” Undeniably, many a clinical physician or expert was biased against measuring diminished nociception in diabetic neuropathy, and for several possible reasons: (1) quantitative sensory testing (QST, see below) was considered inferior to nerve conduction velocity studies in the detection of symptomatic (painful) diabetic neuropathy; (2) Experts of QST not always knew enough of the diabetic foot, this typical complication of symmetrical diabetic neuropathy, to understand what they read. For example those reviewers, who commented on a paper: “In a group of 20 patients (...) the authors compared sensation in ulcerated vs nonulcerated feet and were unable to find any difference. Had a difference been found, one might interpret the sensory dysfunction as a pathophysiological factor in ulcer development”[9]; (3) QST findings were assumed to be poorly reproducible, an argument which was often mixed up with the natural inter-individual variability (on average 5-fold[10]). Intra-individual coefficients of variation are acceptable, for example those of pain perception thresholds are well below 25%[10-14]; and (4) A really important obstacle against routine QST of nociceptive functions was the lack of normative data. However, since 2005, extensive reference material using a particular QST protocol was published[12,14-17], see below. Although this protocol had primarily been devised for examining hyperalgesia in non-nociceptive pain syndromes like fibromyalgia, or neuropathic pains, and not for assessing hypoalgesia[18], the author (Chantelau EA) believes that parts of it may well be applied for studying the failing nociception at the foot in diabetes. Diabetic hyperglycaemia in general does not interfere with QST[19-21]; however, subclinical or clinical hypoglycaemia makes QST virtually impossible.

PLDN: THE PRINCIPAL RISK FACTOR FOR THE DIABETIC FOOT

Peripheral diabetic (poly) neuropathy is the most frequent sequel of diabetes mellitus. The nature of diabetic neuropathy has been clearly established by Martin[22,23] in 1953 and 1954, and by Catterall et al[24] in 1956. While today many people still believe that the condition is always symptomatic with spontaneous pains and dysaesthesias in the legs and feet, these authors had already noted more than half a century ago: “non-myelinated nerve-fibre degeneration occurs in a high proportion of diabetic patients, who, on clinical examination, show no evidence of ‘clinical’ diabetic neuropathy…the presence in diabetics of such nerve-fibre disturbance explains the frequent finding of persistently cold feet and their proneness to traumatic lesions”[24]. Indeed, the basic feature of diabetic neuropathy is small fibre degeneration[25] with reduced or absent nociception at the feet as the key sign (PLDN). By contrast, SDN develops secondary to PLDN, superimposes on PLDN, and affects only a minority of 15%-30% of patients[25,26]. The symptoms of SDN like spontaneous neuropathic pains and tingling in the feet may be alleviated by drug treatment; they resolve gradually while nerve degeneration[27] and PLDN progress. However, diminished nociception in PLDN does not resolve, nor does PLDN respond to drug treatment. Prevalence and incidence rates of PLDN are not known. However, the prevalence of the diabetic foot, indicative of severe, end-stage PLDN, in patients with SDN is about 10%, while the incidence is about 5%-7% per year[28,29].

Most of the patients with PLDN never have complaints until they are facing a foot injury that -unexpectedly - does not hurt. To once more quote Catterall et al[24]: “Cutaneous neuropathic changes commonly start as blisters about the tips of the toes or at the site of a corn or callosity in places constantly exposed to irritation by an ill-fitting shoe. Frequently the deceptive lack of normal sensation, and particular pain sensation, leads a patient to ignore the lesion and delay treatment until secondary infection with deep penetration of the tissues or severe inflammation is present.” PLDN is associated to the duration of diabetes and to the cumulative effects of increased blood glucose (or deficient insulin function). It is symmetrical, age- and length-dependent (it starts in advanced age by affecting the endings of the longest axons in the body, in the skin of the toes[30]). Thus, deficits in nociceptive function develop in the forefeet first, and subsequently ascend to the rearfeet, the ankles and the lower legs. The hands are affected only in most severe cases.

PLDN is a disease of the small primary afferent A-delta and C-fibres: they are thin and their conduction velocity is relatively slow. (A fraction of about 15% of small fibres extending to the skin are efferents; they belong to the sympathetic nervous system and are restricted to the dermis, innervating sweat glands, blood vessels, and the arrector pili muscles. Small fibre efferents will not be discussed here). Primary C-afferents (also named group IV fibres, axon diameter < 1.5 μm, conduction velocity 0.5-2 m/s) are unmyelinated nerve fibres. Primary A-delta afferents (also named group III fibres, axon diameter 1-6 μm, conduction velocity 5-30 m/s) consist of thinly myelinated axons whose intracutaneous endings, however, are unmyelinated[31]. The endings of A-delta and C-afferents do not carry corpuscles but are “free”; they serve as receptors for noxious thermal, mechanical and chemical stimuli (termed nociceptors), and as receptors for innocuous thermal stimuli. In the epidermis they are densely distributed, up to 1000 per 1 cm², which is much more than in muscles, ligaments, joints and bones (periosteum). Most nociceptors are specific of one single modality, while some are polymodal. C-fibre afferents are estimated to account for 70% of all nociceptors in the skin (the remainder are A-delta nociceptors). Cutaneous A-delta nociceptor stimulation results in a short sharp, stinging “first” pain, whereas C-fibre nociceptor stimulation results in prolonged dull, burning “second” pain. In muscles, A-delta and C-fibre nociceptors evoke the same dull pain character[32], see section 4.

In PLDN, the nociceptors degenerate progressively (“die back”) by unknown molecular mechanisms. It is not clear, whether this insidious process of central-peripheral distal axonopathy[33,34] begins already in a prediabetic stage[25]. However, PLDN superimposes on the physiological age-related decline in sensory functions after the age of 50[34-36], and is aggravated by other neuropathic conditions like vitamin B12 deficiency, alcohol toxicity, end-stage renal failure, or exposition to neurotoxic substances or conditions. Clinically, PLDN remains undetected until most nociceptors have died and painlessness of a foot trauma astounds the patient (and the doctor)[24]. The vanishing of intraepidermal free nerve endings can be quantified by histomorphology[31,37,38]; it correlates to increasing severity of diabetic neuropathy[39-41] (Figure 2), and to rising thermal and pain perception thresholds[34,38,40].

Figure 2 Intraepidermal nerve fibre ending density at the lower limb in relation to severity of diabetic neuropathy (adapted from Quattrini et al[41]). HC: Healthy controls; DC: Diabetic controls; Dmi: Mild SDN; Dmo: Moderate SDN; Dsev: Severe SDN; SDN: Symptomatic diabetic neuropathy; INEF: Intraepidermal nerve fibre ending.

A-delta and C- fibre nociceptors and thermal receptors “die back” not only in the skin, but presumably also inside the adjacent subcutaneous tissues (equivalent to group III and IV fibre nociceptors in muscles, fascia, ligaments and joints) of the diabetic foot. However, this remains to be established by histopathology. A-delta (group III) fibres seem to be particularly susceptible to the neuropathic processes, as their functions deteriorate somewhat earlier than C- (group IV) fibre functions[42].

Diabetes affects small fibres prior to large ones. When large myelinated A-beta afferents (axon diameter 6-12 μmol/L, conduction velocity 30-70 m/s) become affected, the axon and the myelin sheet get damaged, and conduction velocity decreases. A-beta fibres mediate touch, deep pressure and vibration sensation by means of their corpuscular endings: Meissner’s, Merkel’s, and Pacinian corpuscles. A-beta fibres are not involved in nociception, but have a role in allodynia[43] which, however, will not be discussed here.

Diabetic foot: End-stage complication of PLDN

The diabetic foot is characterised by painless ulceration and/or arthropathy. Since 60 years or so[22,23,24,44] the condition is well known as an end-stage complication of diabetic neuropathy, of PLDN rather than SDN. The basic defect has frequently been termed “loss of protective sensation”[44], which is, in fact, “loss of nociception” (see below). Ulcers and arthropathy start from single injuries (mechanical, thermal or chemical skin wound, or skeletal trauma) which are subjected to persisting repetitive stress (load bearing by walking, because they do not hurt - Figure 1), whereby they enlarge enormously until they become apparent to the patient. The initial injury may also be caused by uninterrupted repetitive mechanical wear and tear (fatigue from overuse, remaining unperceived because of lack of nociception!) The ulcers become infected, and the infection spreads to the adjacent subcutaneous tissues. Bone and joint damage aggravates due to ongoing unprotected walking. When properly treated immediately (like any other acute foot injury in subjects with preserved nociception!), the initial injuries will heal uneventfully. Human experiments showed that wound healing is not impaired by PLDN[45,46]. Perforating neuropathic foot ulcers heal if they are unloaded completely[24], and if arterial blood flow is sufficient and infection is cured. The ulcers frequently break down again when repetitive stress from walking is resumed, “due to faulty and often excessive weight-bearing by the ulcerated area”[23]. Hence, the overall amputation risk is increased: the diabetic foot is the most frequent cause of amputation in industrialised countries.

CURRENT CLINICAL PRACTICE OF DIAGNOSING DIABETIC NEUROPATHY

According to the common misconception that neuropathy in diabetes in general is SDN, current guidelines state that the condition be diagnosed and staged semi-quantitatively according to various symptom scores (e.g., Neuropathy Symptoms Score, Neuropathy Disability Score, Michigan Neuropathy Screening Instrument, Toronto Clinical Scoring System). However, testing sensory functions is also accepted: five simple nerve tests are considered diagnostic for symptomatic neuropathy, the 10-g monofilament plus 1 of the following 4: vibration using 128-Hz tuning fork, pinprick sensation, ankle reflexes, vibration perception threshold[47]. All except vibration perception threshold are qualitative tests, all but pinprick sensation are unsuitable for diagnosing PLDN.

Vibration and touch represent A-beta fibre functions, not nociceptive functions. Nevertheless, vibration and touch sensation is used as surrogate markers of the risk of sustaining painless foot injuries (ulceration/arthropathy). To identify loss of protective sensation (LOPS) - without assessing nociception- “any of the five tests listed could be used (…). One or more abnormal tests would suggest LOPS, while at least two normal tests (and no abnormal test) would rule out LOPS”[47].

Current guidelines claim that “estimating the severity of diabetic sensorimotor polyneuropathy has not received the attention it deserves. For a given patient with diabetes it is not sufficient to simply identify patients as having sensorimotor polyneuropathy - severity also needs to be ascertained”[48]. However, measuring loss of nociception (LON) as yet has no place in assessing neuropathy in diabetes[47].

Current clinical practice of diagnosing a diabetic foot/diabetic podopathy

Loss of touch or vibration sensation is not diagnostic of an active diabetic foot, while painlessness of a foot injury (soft tissue or skeletal lesion) or infection in a patient with diabetes mellitus is. Three different clinical pictures of active diabetic foot lesions may be discerned, according to the dominant clinical component: (1) septic; (2) ischaemic; and (3) arthropathic. After healing of such a lesion, the foot is an inactive diabetic foot (with scars, deformities, etc.) and remains a locus minoris resistentiae for the rest of the life. LON traditionally is not assessed systematically, but is taken as matter of fact when a patient with an injured foot does not limp. The patient history of traumatic events is often unproductive as far as normal symptomatology (pain!) is concerned. However, when concurrent symptoms of trauma onset other than pain are concerned, like swelling and erythema, patient history may be very well productive.

QST

The QST protocol, published by the German Research Network on Neuropathic Pain and supported by pharmaceutical companies, was designed for non-invasively diagnosing pain syndromes like spontaneous neuropathic pains, fibromyalgia or chronic lower back pain[15,16,18]. It comprises 13 somatosensory sensory tests that measure the function of large (A-beta) and small (A-delta and C-) afferent nerve fibres, and the corresponding functions of the spinothalamic tract (cold detection threshold, warm detection threshold, thermal sensory limen, paradoxical heat sensation, cold pain threshold, heat pain threshold, mechanical detection threshold, mechanical pain threshold, mechanical pain sensitivity, dynamic mechanical allodynia, wind-up ratio, vibration detection threshold, and deep pressure pain threshold). Using this protocol in healthy subjects, a pool of normative reference data from face, hand and foot had been generated[16,17].

QST of pain perception in diabetes

Of all sensory modalities that are impaired by diabetes mellitus, the nociceptive pain system is the most relevant. In the following sections, the focus will therefore be directed on pain perception thresholds: elevated thresholds indicate reduced perception of noxious stimuli, and vice versa. Pain tolerance or other sensory modalities (vibration, electrical current, chemicals, touch, thermal perception) will not be addressed. Some technical features of the threshold studies considered here are outlined below, concerning thermal (cutaneous heat and cold), and mechanical (cutaneous and deep pressure) pain perception.

Cutaneous thermal pain perception: Thermal pain perception threshold is measured by applying warm or cold stimuli to the skin, generated by appropriate thermodes (Figure 3). The temperature of the thermode is gradually increased or decreased. The subject under study is asked to indicate whether a thermal sensation becomes painful. Heat pain represents C-fibre function, and cold pain represents A-delta fibre function. In healthy subjects, heat pain perception threshold may range from approximately 41 °C to approximately 47 °C (average approximately 44 °C), and cold pain perception threshold ranges from approximately 1 °C to approximately 30 °C (average approximately 12 °C).

Figure 3 MSA Thermotest^® (electronic device, expensive!).

Deep pressure pain perception: Deep pressure pain perception threshold (DPPPT) may be examined, for instance, by a hand held device (algometer, Figure 4) containing a mechanical or electronic force gauge, and a plunger with a flat blunt rubber tip. The diameter of the tip can be 11 mm (surface approximately 1 cm²), or lager, as appropriate. The plunger is pressed perpendicularly onto the skin and the underlying structures, with slowly increasing stimulus ramp (50 kPa/s, equivalent to 0.5 kg/s). The subject under study communicates when his pressure sensation turns to pain.

Figure 4 Algometer^® (electronic device, expensive!).

The deep pressure pain sensation evoked by an algometer at muscle or joint is as yet not fully understood. For instance, the afferents involved (high and low threshold mechanosensitive units[32]), the anatomical structures subjected to algometer pressure, and the role of spatial summation (pressure area 1 cm²)[49,50], still have to be elucidated. The contribution of fascia, which is more pain-sensitive than muscle, and of periosteum, which is more pain-sensitive than bone marrow, remains to be determined. The pain quality at reaching DPPPT is dull and burning and probably more of a pressure discomfort than of a stinging or pricking[49,50]. Inside deep tissues, stimulation of the few A-delta (group III) nociceptors simultaneously probably does not elicit a separate pricking pain sensation like a single punctuate stimulation at the skin does[49,51]. DPPPT decreases at chronically ischaemic muscle[52,53]. DPPPT at muscle (blunt stimulation) decreases with age, at variance to cutaneous pressure pain perception threshold (CPPPT, punctuate stimulation)[42]. Using a pressure algometer with 1 cm² contact area to stimulate musculus abductor hallucis, DPPPT in healthy subjects may range from approximately 200 kPa to approximately 1000 kPa (average approximately 500 kPa, equivalent to 5 kg). Algometer stimulation of a joint evokes a DPPPT of the same range.

Cutaneous pressure pain perception: Cutaneous pressure pain perception threshold (CPPPT) is examined by the use of calibrated pinprick stimulators with a sharp edge, comparable to von Frey hairs. These punctuate stimulators are filaments of 50-350 μm diameter, with calibrated bending forces (or weight loads) over a range from 8 mN to 512 mN (equivalent to 0.8 g to 51.2 g) (Figure 5).

Figure 5 MARSTOCKnervtest^® PinPrick stimulator 512 mN (fibreglass, cheap!).

The investigator presses each single stimulator perpendicular to the skin (until bending of the filament) and the subject under study communicates, whether a painless touch is felt, or a painful (pricking or stinging) sensation, or nothing at all. The pain threshold is determined by the method of limits, or by the forced-choice-method. The stimulators excite single nociceptors representing A-delta (rather than C-fibre) functions; the pain quality evoked is that of a stinging, pricking “first” pain. CPPPT increases with age (at variance to DPPPT, see above). At the non-callous plantar skin of healthy subjects, CPPPT may range from approximately 4 mN to approximately 450 mN (average approximately 100 mN, equivalent to 10 g).

Limitations of pain threshold measurements: QST is a subjective psychophysical method, as it is based on the patient report. “A critical element in pain threshold determination is the particular sensory experience an individual considers painful. This factor will be influenced by, among other things, the subject’s pain experience history, and the instructions given by the experimenter. For instance, thresholds are likely to be different if a subject is instructed to indicate when he perceives ‘pain’vs when he perceives ‘a sharp or burning sensation’ or ‘an uncomfortable situation’….”[54] In particular, pressure pain measurement is affected by the hardness of the skin at the site of measurement, and by the ramp rate of increasing the stimulus of hand-held algometers, amongst others.

QST of nociception at the diabetic foot

Early studies of pain perception thresholds: Between 1988 and 2011, not a single study could be identified on cold pain perception threshold, or on DPPPT. However, there are at least 4 studies on CPPPT and/or heat pain perception threshold[29,55-57], all of which found significantly elevated thresholds at the active (?) diabetic foot and the contralateral foot, as compared to neuropathic patients without diabetic foot (Table 1). In neuropathic patients with and without diabetic foot, pain perception thresholds were significantly higher than in non-neuropathic (healthy or diabetic) control subjects. Lang et al[52] reported that perception thresholds for cold pain, heat pain and pinprick were elevated at legs with chronic severe foot ischaemia (rest pain with/without tissue loss); the 16 patients (8 of whom had diabetes mellitus) had polyneuropathy according to a vibration perception threshold < 4/8 (graduated tuning fork). DPPPT at ischaemic muscle, however, was lower than at healthy controls’ muscle (199 kPa vs 295 kPa[52]). In another study of non-neuropathic patients with chronic leg ischaemia, CPPPT, and DPPPT at muscle, were lower than in control subjects but cutaneous cold and heat pain perception thresholds were not[53].

Table 1 Early studies of cutaneous pain perception thresholds (pressure, heat) at the diabetic foot, using various, non-standardised methods[29,55,56,57].

Test	Diabetic subjects under study		Healthy	Ref.
	Diabetic foot	Neuropathy	controls
CPPPT	1001	100	0	[29]
CPPPT	1001	631	-	[55]
CPPPT	68	30	2	[56]
Heat	1001	100	0	[29]
Heat	1001	100	0	[57]

Percentages of supranormal results are shown.

¹Above upper limit of measurement. CPPPT: Cutaneous pressure pain perception threshold.

Recent pilot studies of pressure pain perception thresholds: Between 2012 and 2014, we conducted four studies on the matter[58-61] including explorative cross-sectional and follow-up studies. Our overall database comprised 123 subjects in whom pressure pain perception thresholds were measured: 43 control subjects, 23 of whom had an acute trauma of the foot skeleton (sprain, minor fracture), 59 diabetic patients with PLDN (46 of whom with a diabetic foot), and 21 diabetic patients without PLDN. Among the 46 patients with a diabetic foot, there were 33 with active painless foot ulcer, 13 with healed foot ulcer, 11 with healed Charcot foot (grade 1, inactive[62]), and 13 with acute skeletal trauma (elective surgery). CPPPT and DPPPT were studied, as well as vibration perception threshold (not reported here). According to our study protocol, CPPPT was measured only at the forefoot (plantar digital skinfold), while DPPPT was measured at the hindfoot (m. abductor hallucis) and at the forefoot (metatarsophalangeal joint) (Figure 6).

Figure 6 Pressure pain perception threshold measurements done in own studies[61]. CPPPT: Cutaneous pressure pain perception threshold; DPPPT: Deep pressure pain perception threshold.

By cross-sectional comparison, CPPPT at a plantar digital skinfold (glabrous skin) was above the safety limit of measurement of 512 mN (approximately 51.2 g) in 98% of cases with a past or present painless foot ulcer or healed arthropathy (Charcot foot). This is in line with the older reports cited above. Simultaneously, DPPPT at abductor hallucis muscle was above the safety limit of measurement (1400 kPa, equivalent to 14 kg) only in about 20% of cases. The DPPPT at metatarsaophalangeal joint was above 1400 kPa in about 50% of cases (Table 2).

Table 2 Simultaneous measurements of pressure pain perception thresholds at the foot in subjects with and without painless diabetic neuropathy, with and without diabetic foot.

Subject populations	Diabetic foot1	Healthy2	PLDN3	DM-controls4
No. of subjects	46	43	13	21
% of subjects with CPPPT > 512 mN at toe skinfold	98%	4%	46%	5%
% of subjects with DPPPT > 1400 kPa at m. abductor hallucis	22%	0%	0%	0%
% of subjects with DPPPT > 1400 kPa at metatarsophalangeal joint	50%	2%	0%	5%

¹With previous or active ulcer/arthropathy;

²Non-neuropathic, healthy;

³PLDN, no ulcer;

⁴Diabetes, no neuropathy, no ulcer. Percent subjects with thresholds above the safety limit of measurement are shown. Adapted from ref. [58-60]. CPPPT: Cutaneous pressure pain perception threshold (512 mN upper limit of measurement); DPPPT: Deep pressure pain perception threshold (1400 kPa upper limit of measurement).

Serial pressure nociception studies to monitor pain thresholds after an acute foot trauma (ankle sprain, toe fracture) in otherwise healthy subjects revealed that CPPPT and DPPPT at the site of the trauma decreased slightly, indicating normal, inflammation-mediated posttraumatic hyperalgesia (Figure 7). This observation is consistent with previous reports (Martinez et al[63]: decrements in CPPPT on day 1-4 at an operated knee; Lang et al[53]: reduction in CPPPT and DPPPT in a chronic ischaemic leg up to 3 mo after successful endovascular revascularisation, Dominguez et al[64]: reduced DPPPT at the foot immediately after a sprain). This physiologic, inflammation-mediated posttraumatic hyperalgesia is the essential trigger of avoidance behaviour to preserve the injured limb from further noxious and innocuous impacts.

Figure 7 Lowered pain perception thresholds (hyperalgesia) after acute skeletal trauma of the foot in 13 healthy control subjects (sprain, toe fracture) and in 12 diabetic foot patients (elective surgery) at the injured site. Data are combined from ref. [59,61]. A: Cutaneous pressure pain perception threshold (mN) (pinprick) at plantar digital skinfold (medians); B: Deep pressure pain perception threshold (kPa) over metatarsophalangeal joint (medians); C: Deep pressure pain perception threshold (kPa) over musculus abductor hallucis (medians).

By contrast, at the diabetic foot CPPPT was extremely elevated (above the safety limit of measurement) prior to an acute foot trauma (elective surgery), and also immediately thereafter (Figure 7A); there was no posttraumatic decrement of CPPPT at the site of the trauma[61]. This suggests absence of A-delta “first” pain quality and inflammation-mediated hyperalgesia. DPPPT at the diabetic foot, being elevated prior to the surgery, showed a small posttraumatic decrement from this baseline level (Figure 7B and C) which, however, was not accompanied by clinical signs of hyperalgesia, or avoidance reaction, respectively.

These data, however, have to be interpreted with caution since our study groups were small (only 12 and 13 subjects, respectively) and heterogeneous. Furthermore, it cannot be ruled out that our study subjects may have behaved like those young healthy women, who reacted to consecutive daily measurements of DPPPT over 4 d by increased pain sensitivity[65] - a reaction that may be an exception rather than the rule[13]. CPPPT within the normal reference range may decrease slightly upon re-testing[14] but remains consistently above the safety limit of measurement in patients with diabetic foot[58-61].

Miscellaneous studies of A-delta and C-fibre sensory functions at the diabetic foot: At least 7 studies[66-72] could be ascertained, all of which reporting that the axon-reflex reaction is not much smaller in neuropathic patients with a diabetic foot than in those without. Axon-reflex reaction in the feet of neuropathic patients was consistently smaller than in the feet of non-neuropathic control subjects.

At least 11 studies[57,67,73-81] could be ascertained, all of which reporting that neuropathic patients with a diabetic foot have higher thermal detection thresholds than neuropathic patients without. Thresholds were higher in the feet of neuropathic patients than of control subjects, and were particularly high in patients with more severe, relapsing diabetic podopathy and/or Charcot foot[76-79]. One study[79] showed a lower cooling detection threshold in the active or inactive Charcot foot vs the contralateral foot, while the warming detection threshold was similar on both feet.

CONCLUSION

At the diabetic foot, cutaneous thermal and mechanical (pinprick) nociception is reduced dramatically. CPPPT was elevated beyond the safety limit of measurement in most of the small number of published studies. Patients with diabetic foot ulcers unanimously display a CPPPT at a digital plantar skinfold which is beyond the safety limit of 512 mN. However, details of the causal role of increased CPPPT (indicating lack of A-delta nociception) for the manifestation of diabetic podopathy remain to be established, as well as the anatomical distribution of PLDN at the foot. Concerning DPPPT at the diabetic foot, the data is less clear, as differences to control subjects often were surprisingly small. Compared to baseline DPPPT, inflammation-mediated posttraumatic DPPPT was reduced at the diabetic foot, albeit at a high level and far from indicating hyperalgesia. Pending confirmation, this may suggest that DPPPT probably does not need to be extremely elevated for a diabetic foot to develop.

There are many more open questions, for instance: what number of residual intraepidermal free nerve fibre endings corresponds to the elevation of CPPPT at the diabetic foot? What elevation of CPPPT is necessary to appreciably increase a patient’s risk of sustaining painless foot ulceration and/or arthropathy? How much pain threshold elevation, i.e., how much “LON”, is clinically meaningful? Is the nociception at healthy tissues relevant, or is it the pain perception (hyperalgesia, allodynia), that is typically caused by stimulation of inflamed tissue? Is superficial and deep tissue nociception similarly impaired by PLDN? How much are “silent” nociceptors[82] affected? Is the stinging “first” pain more reduced than the burning “second” pain? Is there a defect in posttraumatic ongoing (non-stimulated) pain? Is absolute or relative reduction of individual pain perception essential for the diabetic foot to develop? Is pain perception or pain tolerance more relevant? Do abnormalities of the central nervous system contribute to the painlessness of the diabetic foot? To what extent is sensitization, either peripheral or central, affected at the diabetic foot? Concerning the discrepancies between CPPPT (at the forefoot) and DPPPT (at muscle of hindfoot) in our own studies[58-61] - does the distal-to-proximal gradient of PLDN provide an explanation?

Measuring nociception at the diabetics' feet-what is it good for?

LON is the principal pathogenic component of diabetic podopathy, as it is the principal functional component of PLDN. Loss of pressure perception does not increase the risk of diabetic foot ulceration, whereas a history of foot ulceration (equivalent to cutaneous LON) does[83]. Measuring nociception provides the opportunity for a clinically meaningful staging of the severity of PLDN. A critical LON needs to be elaborated (probably corresponding to a critical number of intraepidermal nerve fibre endings), “to be able to inform the patient adequately about the risk of ulcer formation and to prescribe preventative measures… otherwise, sensation loss and resultant ulcers may lead to amputation”, as Assal et al[84] advocated 25 years ago. Perhaps every patient may have his/her own individual critical LON to be taken into account. Small reductions in intraepidermal nerve fibre density may be mirrored by respective increases in CPPPT, as recent work by Selim et al[85] shows. Hence, measuring CPPPT prospectively may be suitable for monitoring progression of PLDN.

Shifting the emphasis from amputation prevention by wound healing to amputation prevention by injury prevention: Once an active diabetic foot/podopathy has become manifest, a multitude of cumbersome, laborious, tedious, expensive and costly therapeutic interventions is necessary to save the foot and prevent an amputation. Injury prevention in high-risk PLDN - to prevent the diabetic foot- is certainly much more efficient. Preventing the first or recurrent ulcers[86] should be no question. Having had a first ulcer strongly predisposes to subsequent ones, most likely due to comorbidities from structural abnormalities that resulted from the (healed) first ulcer (scars, deformities, contractures, tissue loss). The same holds true for arthropathy. Hence, preventing the first ulcer/arthropathy must have absolute priority. To this end, intensive prophylactic measures are necessary in high-risk patients who should be identifiable by simple means. As a proposal, the following tentative risk-stratification is construed from various published data (assuming a “normal” average walking activity of 5000 strides per day [87]); (Table 3).

Table 3 Percentages of subjects with active diabetic podopathy (ulcer, arthropathy) according to loss of nociception.

Foot morbidity	Degree of loss of nociception			Ref.
	None	Mild/moderate	Severe
	(CPPPT1 approximately 100 mN)	(CPPPT1 < 512 mN)	(CPPPT1 > 512 mN)	[59]
Past/present painless ulcer	0%	0%	100%	[28,59,77,86-101]
Foot ulcer, incidence per year	Approximately 0%-1%	Approximately 6%-15%	Approximately 20%-100%2	[28,77,86-101]
Charcot-foot, prevalence	Approximately 0%	Approximately 0.05%	Approximately 2%	[102]

¹Cutaneous pressure pain perception threshold, pinprick stimulator, upper limit of measurement 512 mN, measured at a plantar digital skinfold;

²Depending on compliance with prophylactic podiatry and/or special footwear. CPPPT: Cutaneous pressure pain perception threshold.

Subjects with critical LON at their feet require signals other than foot pain to control foot usage and avoid over-usage. These signals can only come from technical devices, like step counters or load monitors incorporated into footwear, that send acoustic or optical alarms when a critical load (or an equivalent number of steps) has accumulated. Then, the feet have to be rested - for an appropriate period of time as physiologically required to regenerate the stressed tissues - before their usage (walking) may be resumed. Moreover, feet with critical LON need special footwear to avoid rubbing of the skin while walking in order to prevent skin abrasion and blister formation.

Pain is a homeostatic emotion that drives behaviour, just as hunger and thirst are. Patients who cannot have it, either from acquired or inherited insensitivity to pain[103,104], either in the feet or elsewhere in the body, need our special protection, and much more specific than we, their caregivers, may have been willing to concede in the past. Research into nociception at the foot in diabetes must continue and expand. Chances to improve injury prevention will be missed, if this uncharted territory remains unexplored.