Sensitivities to Thermal and Mechanical Stimuli: Adults with Sickle Cell Disease Compared to Healthy, Pain-free African American Controls

Abstract

Evidence supports^, but is inconclusive that sensitization contributes to chronic pain in some adults with sickle cell disease (SCD). We determined the prevalence of pain sensitization among adults with SCD pain compared with pain-free healthy adults. In a cross sectional, single session study of 186 African American outpatients with SCD pain (age 18–74 years, 59% female) and 124 healthy age, gender, and race matched control subjects (age 18–69 years, 49% female), we compared responses to standard thermal (Medoc TSA II) and mechanical stimuli (von Frey filaments). Although we observed no significant differences in thermal thresholds between controls and patients, patients with SCD had lower pain thresholds to mechanical stimuli and reported higher pain intensity scores to all thermal and mechanical stimuli at a non-painful body site. Compared with controls, about twice as many patients with SCD showed sensitization: 12% versus 23% at the anterior forearm site (p=.02), and 16% versus 32% across three tested sites (p=.004). Among patients with SCD, 18% exhibited some element of central sensitization. Findings indicate that persistent allodynia and hyperalgesia can be part of the SCD pain experience and should be considered when selecting therapies for SCD pain.

Info graphic

Density plots for healthy controls and patients with SCD showing pain intensity values assigned to thermal (cold, heat) and mechanical pain thresholds.

Color key for level: Lighter regions represent higher densities, namely more patients with those measurement values.

Summary

Compared with 124 age-, sex-, and race-matched healthy controls, 186 patients with pain and sickle cell disease (SCD) showed sensory sensitivity to thermal and mechanical quantitative sensory testing. Among the 32% of patients with SCD and sensitization, more than half exhibited some element of central sensitization.

Introduction

Until recent findings that sickle cell disease (SCD) pain has characteristics consistent with central or peripheral sensitization,^{12, 27, 28, 43} it was typically viewed as acute or persistent pain and treated mostly with opioids. Often the etiology of SCD pain is thought to be only episodic, driven by vaso-occlusion and somatic or visceral tissue damage,^{1, 2, 4} however, it has become increasingly clear that adults with SCD also experience chronic pain. A growing body of evidence now supports,^{12, 27, 28, 40, 43,} but is not conclusive, that sensitization contributes to chronic pain in some adults with SCD.^{36, 43} The study purpose was to examine the prevalence of pain sensitization among adults with pain and diagnosed with SCD compared with a matched group of pain-free healthy adults.

Altered ascending sensory processing within the peripheral or central nervous systems (peripheral and central sensitization) may contribute to the chronic pain in patients with SCD, likely as a result of repeated vaso-occlusive episodes and tissue ischemia over their entire lives.⁴⁵ Characterizing the involvement of peripheral and central sensitization to SCD pain may allow for the use of more effective treatments than opioids. Using a multidimensional pain tool that measures pain location, intensity, quality, and pattern,⁴³ we found that 92% of 145 adults with SCD selected descriptors commonly reported by people with sensitization-related pain (e.g., burning, cold, numb, radiating, shooting, and stabbing).^{12, 43} Additionally, using thermal and mechanical quantitative sensory testing (QST), we found that 24/25 adults with SCD had allodynia or hyperalgesia to cold, heat and mechanical stimuli, which implicates central, peripheral, or mixed sensitization.¹² Although these findings are intriguing, they were limited by reliance on literature-based QST norms that were not sufficiently specific to the patients’ pain sites or the age range of the adults included in the sample. Campbell et al.⁶ subsequently reported that 17 SCD patients with low central sensitization compared to 21 SCD patients with high central sensitization based on QST temporal summation parameters differed significantly on a variety of QST measures and patient outcomes including pain, opioid use, psychological, and acute care utilization characteristics. Campbell et al.⁷ also found that 83 SCD patients differed from 27 healthy controls on pressure pain threshold, heat pain tolerance, and hot water hand immersion tests. They also found that 27 SCD patients differed from 27 health controls on inflammatory markers.⁷ Others^{6, 16, 17, 21, 30} noted sensitivity to pain stimuli among children and adolescents with SCD (N=35–64) including sensitization on some QST parameters compared to age-matched healthy controls.^{6, 21}

Consistent with these findings of increased sensitization associated with neuropathic pain in people with SCD pain, several groups have reported mechanisms of neuropathic pain in transgenic sickle cell mouse models.^{15, 18–20, 32, 33, 47} Ongoing spontaneous pain, cold allodynia, mechanical allodynia, and heat hyperalgesia were observed in two transgenic mouse models of SCD.^{18, 19} In these mouse models, a CaMKIIα mechanism was implicated as underlying the chronic SCD pain and CaMKIIα inhibitors were effective in reducing the ongoing spontaneous pain, cold allodynia, mechanical allodynia, and heat hyperalgesia.^{18, 19} In a translational safety study, we showed promise for pain relief from a CaMKIIα inhibitor in 18 adults with SCD.²⁸

These findings from studies with small sample sizes are intriguing. The sensitization phenomenon requires examination in a larger SCD cohort with pain and matched healthy controls who complete the same QST protocol focused on sensory functions that would identify presence of allodynia and hyperalgesia. The specific aim was to compare the prevalence of pain sensitization in adults with SCD versus sex-, age- and race-matched healthy control adults using mechanical and thermal stimulation of sensory fibers (Aβ [mechanical, von Frey QST], A-δ [cold QST], C [heat QST]). We hypothesized that a significantly higher proportion of the SCD sample would report allodynia/hyperalgesia indicative of sensitization than the healthy control sample.

Methods:

Study Design

As part of a larger project completed in February 2019, we conducted a cross-sectional, single session study of African American outpatient adults with SCD and healthy control subjects matched on age, sex, and race to compare their responses to standard thermal and mechanical stimuli. The study was approved by the Institutional Review Board at the University of Illinois at Chicago (UIC).

Setting and Sample

We recruited potential participants in the community or at the University of Illinois Hospital Sickle Cell Clinic. Community recruitment strategies included the use of online advertisement and social media like Craigslist and Facebook. Craigslist was particularly successful in the recruitment of the healthy control subjects. Many healthy control subjects were also recruited through word of mouth, referrals from SCD patients, and team members’ acquaintances. We used attractive and succinct flyers to aid recruitment. The UIC was the data collection site.

The study’s inclusion criteria required that the healthy control subjects: (a) were of African ancestry, (b) had a health history negative for diabetes, hypertension, SCD, cancer, current acute or chronic pain, (c) spoke and read English; and (d) were 18 years of age or older. Volunteers were excluded if they were: (a) legally blind, (b) unable physically to complete study measures, or (c) reported using prescription pain medications or recreational drugs.

The eligibility criteria for the patients were: (a) diagnosis of SCD, (b) pain (≦3 on 0–10 scale) related to SCD within the previous 12 months, (c) chronic SCD pain, defined as patient reported SCD pain > 0 on at least one half of the days during the previous 3 months, (d) spoke and read English, (e) 18 years of age or older, and (f) African ancestry. Patients were excluded if they were: (a) legally blind, (b) physically unable to complete study measures, or (c) diagnosed with a polyneuropathy or diabetes mellitus recorded in their medical records. We did not exclude patients using opioid, non-opioid, or adjuvant pain medicines.

We approached a total of 354 patients with SCD between March 2015 and August 2017. Fifteen were excluded for no or low pain, 9 for diabetes, and 5 miscellaneous reasons. Of the 325 eligible, 43 were not interested, 96 could not be scheduled and 186 enrolled and completed study measures.

Procedures

Coauthor (RBF) shared his protocol for thermal detection and pain threshold measurements and conducted a training session for the research specialists and other research team members. We trained the research specialists in all study procedures and monitored the fidelity of their performance over the course of the study. After the research specialist obtained the participant’s informed consent, the research specialist trained the participant in all study procedures. Then the participant completed the standardized QST procedures for thermal and mechanical stimuli. The testing protocol was adapted from previously tested protocols¹² and consistent with the EFNS (European Federation of Neurological Societies) recommendations for testing Aβ, A-delta, and C fiber function.¹¹ If the thermal or mechanical stimulus was perceived as painful, participants verbally reported that pain intensity using the Pain Intensity Number Scale (PINS).⁴¹

Measures

PINS Measure

The PINS measured the patient’s perception of the level of pain now⁴¹ as ratio level data.²⁹ The patient designated the pain as a number between 0 and 10, where 0 is “no pain” and 10 is “pain as bad as it could be.” Standardized instructions,²⁹ concurrent (r=.80 to .89) and construct²² validity have been reported. Reliability and sensitivity of the PINS have been studied less well.²⁹ In a cancer sample, PINS measures separated by two weeks were correlated at a moderate level (r₍₈₇₎=.51, p<.001).⁴⁴ Patients with cancer complete the PINS in less than 1 minute, but several minutes may be needed for instruction and practice the first time it is used.⁴¹ Use of a standardized practice session helps the very young,³⁸ the elderly,⁴² and adults with SCD⁴³ to use the PINS to report pain intensity.

QST Thermal Measures

We completed the thermal QST using a Medoc TSA-II NeuroSensory Analyzer, a computer controlled instrument that uses a Peltier device to deliver highly repeatable thermal stimuli. The TSA-II permits quantitative assessment of small-caliber (A-δ and C fiber) sensory nerve function, and is used to identify thermal pain thresholds. We used the 30 mm × 30 mm probe and the methods of limits protocol without overlapping test areas within each test site.

For each participant (healthy controls and patients), testing was done at three sites, selected from the following six priority test sites: anterior forearm, posterior forearm, lateral upper arm, lateral leg, medial leg, and posterior leg. For the patients, we tested two painful sites and then one non-painful (reference) site. The patients reported the location of their pain on the PAINReportIt body outline, and two painful sites for each patient were selected based on close proximity to the six priority areas. For the controls, all the tested sites were non-painful per our inclusion criteria.³¹ If possible, we used the anterior forearm as the reference site for all participants (non-painful site for patients). For each participant, we tried to include at least one upper and one lower extremity site. Each participant was trained in the procedures and completed a practice trial first so that they could become familiar with the instrument and the temperature change of the instrument’s thermode. We always used the reference site for the practice trial. For the actual trial, the reference site was tested last to allow sufficient time (30 ± 10 minutes on average) between practice and test on the same site. For patients, we tested pain on the opposite side of the body as the non-painful site, if possible, to obtain valid tests of central and peripheral pain mechanisms.

We conducted the study in a private, temperature controlled research room and followed the ascending methods of limits protocol. After placing the TSA-II thermode on the participant’s skin, the temperature increased from a baseline of 32° C at a 0.5° C/second (for warm sensation or heat pain) or decreased from a baseline of 32° C at a 0.5° C/second (for cool sensation or cold pain). The participant responded to the temperature stimuli by pressing a button, indicating detection of the cool or warm sensation and first perception of pain as a result of the cold or hot stimulus, at which point the thermode returned to the baseline temperature. We conducted from three to five repetitions of each measure at each site, with the goal of obtaining a set of three measurements within 2° C of each other. We stopped at five repetitions even if the difference was larger than 2° C. The three closest measurements were averaged to obtain the cool detection threshold (CD), warm detection threshold (WD), cold pain threshold (CPTh), and heat pain threshold (HPTh). The corresponding PINS scores in response to the cold and heat stimuli were also recorded as cold pain intensity (CP) and heat pain intensity (HP). To avoid sensory summation at any one location, we completed the repetitions in adjacent but non-overlapping areas at each of the test sites.

QST Mechanical Measures

We conducted assessments for mechanical detection (MD) and pain threshold (MPTh) using standardized, calibrated von Frey filaments. These are filaments of different thicknesses that are calibrated to bend at a set amount of force in grams (g), depending on the thickness. When a calibrated filament is applied perpendicularly to the area being tested, this punctate mechanical stimulus produces a non-painful or a painful percept. Consistent with the EFNS protocol,¹¹ we used seven filaments: 3.84 (0.6g), 4.17 (1.4g), 4.56 (4.0g), 4.74 (6.0g), 5.07 (10g), 5.46 (26g), 5.88 (60g). For each successively thicker filament, the participant reported if the filament was felt, if so, if it was painful, and if painful, the level of pain on the 0–10 PINS. We tested the same sites used for the thermal testing and conducted up to three repetitions of each filament, applying successively thicker filaments until a filament was reported as painful, at which point the MPTh and the corresponding mechanical pain intensity (MP) were recorded.

Analysis

Data collected from quantitative sensory testing were imported into statistical software R for analysis. Descriptive statistics (frequencies and percentages) of the demographic variables were obtained. Chi-square test and Fisher’s test were used for comparing the demographics of controls and patients. There was minimal (<0.5%) missing data and complete-case analysis was performed. To detect the sensitization of the A-δ fibers (cold sensation), C fibers (heat sensation), and Aβ fibers (mechanical sensation), QST thermal and mechanical pain thresholds and pain intensity levels were compared to pre-defined thresholds (Table 1). We focused on participants exhibiting extreme sensitivity to stimuli. More specifically, sensitization of Aβ fibers was defined as pain reported at the lowest pressure (0.6g) or moderately severe pain (PINS scores 6 or higher) in response to other pressure stimuli; thermal sensitization as pain reported within 0.5° C of baseline (32° C) for upper body sites or within 1.5° C of baseline for lower body sites based on previous findings³¹ or moderately severe PINS scores (6 or higher) in response to the thermal stimuli (Table 1).

Table 1.

Criteria for sensitization by fiber and quantitative sensory testing parameter.

Test Site	A-delta sensitization	C sensitization	A-beta sensitization
Lower body	CPTh≥30.5° C or CP≥6	HPTh≤33.5° C or HP≥6	MPTh=0.6g or MP≥6
Upper body	CPTh≥31.5° C or CP≥6	HPTh≤32.5° C or HP≥6	MPTh=0.6g or MP≥6

Key: CPTh = cold pain threshold, HPTh = heat pain threshold, MPTh = mechanical pain threshold; CP = cold pain intensity on 0–10 scale, HP = heat pain intensity on 0–10 scale, MP = mechanical pain intensity on 0–10 scale

Consistent with the International Association for the Study of Pain’s (IASP) definition of sensitization, we defined sensitization from a clinical perspective as allodynia (“pain due to a stimulus that does not normally provoke pain”) or hyperalgesia (“increased pain from a stimulus that normally provokes pain”).³⁷ Central sensitization mechanisms are implicated if the allodynia and hyperalgesia extend beyond the receptive fields of neurons innervating the painful area (widespread allodynia/hyperalgesia) such as pain from light touch at the non-painful test site.³⁵ Peripheral sensitization mechanisms are implicated if the allodynia and hyperalgesia are restricted to the receptive fields of neurons innervating the painful area (localized allodynia/hyperalgesia of Aδ and C fiber stimuli).³⁵

As displayed in Figure 1 and based on basic pain mechanisms,^{13, 39} if the participant was sensate to the stimuli and all QST findings were negative for allodynia or hyperalgesia at all sites, then the participant was classified as having no sensitization. If the non-painful site Aβ findings (von Frey filaments) were positive for allodynia or hyperalgesia, then the participant was classified as having central sensitization. If, on the other hand, the non-painful site findings were all negative and A-δ or C findings (thermal) were positive at both painful sites, then the subject was classified as having peripheral sensitization. Otherwise, the participant was classified as having mixed sensitization. The decision tree in Figure 1 allows integration of all the thermal and mechanical QST findings across the three test sites. The percentages of participants with signs of sensitization among healthy controls and among patients with pain and SCD were compared using Chi-square tests and using binary logistic regression controlling for sex and age group.

Figure 1.

Decision Tree for Categorizing Each Subject as Having Sensitization Based on Thermal and Mechanical Quantitative Sensory Testing Across the Reference Test Site and Two Painful Test Sites. © 2013, 2018: D. J. Wilkie, Y. Yao, R. E. Molokie, & Z. J. Wang, used with permission of these listed copyright holders.

Results

Description of sociodemographic and study variables

The demographics of the 124 healthy controls (hereafter, called controls) and 186 patients with SCD and pain (hereafter, called patients) in our study are presented in Table 2. Group differences in sex, age, race, ethnicity, or marital status were not statistically significant. The groups differed significantly in their education and income levels (p<.001 and p=.001, respectively). The control group included a larger proportion of participants with bachelor’s degrees and higher income levels than the SCD group. In Table 2, we report the distribution of the controls and patients in two age groups (18 to 39 vs 40+ years). The mean ages of the controls (38.6±12.5) and patients (36.6±11.7) did not differ significantly (p=.16). Patients reported an average pain intensity of 4.5±2.4 and pain located at an average of 7.0±4.1 body sites. Most (57%) patients had at least 4 pain crisis episodes in the past 12 months, followed by 30% that reported 2–3 episodes, and a small percentage (13%) that reported 0–1 episode. Most patients had SS genotype (73%), followed by SC (19%) and other (8%) genotypes. Their history of stroke or hydroxyurea use was not documented.

Table 2.

Demographic characteristics of the two samples: healthy controls and patients with SCD.

Variable	Category	Controls (n=124)	Patients (n=186)	p value
Sex	Female	61 (49%)	110 (59%)	.11
	Male	63 (51%)	76 (41%)
Age	18–39	64 (52%)	118 (63%)	.05
	40+	60 (48%)	68 (37%)
Race	Black	118 (95%)	180 (97%)	.71
	Mixed	5 (4%)	4 (2%)
	Other	1 (1%)	2 (1%)
Ethnicity	Hispanic	4 (3%)	5 (3%)	1
	non-Hispanic	120 (97%)	181 (97%)
Marital Status* (missing: 3 ctrl; 4 SCD)	Single	88 (73%)	130 (71%)	.91
	Married/Partnered	29 (24%)	43 (24%)
	Divorced/Separated	2 (2%)	6 (3%)
	Widowed	2 (2%)	3 (2%)
Education (missing: 2 ctrl; 4 SCD)	High school or less	57 (47%)	79 (43%)	<.001
	Some college	34 (28%)	84 (46%)
	College or higher	31 (25%)	19 (10%)
Income (missing: 5 ctrl; 10 SCD)	$0–$10,000	35 (29%)	74 (42%)	.001
	$11,000–$30,000	38 (32%)	69 (39%)
	$30,000+	46 (39%)	33 (19%)

^*For those choosing other but indicating that they were in a relationship, we put them in the Married/Partnered category; ctrl = healthy controls; SCD = sickle cell disease

Descriptive statistics for pain thresholds and pain intensity levels in response to thermal and mechanical stimuli across all demographic groups are given in Table 3 for the reference test site (forearm anterior for control; non-painful site for patients). The differences between controls and patients in CPTh and HPTh were not statistically significant. Patients reported lower MPTh and higher CP, HP, and MP than controls.

Table 3.

Descriptive statistics for thresholds for the reference test site (usually anterior forearm, non-painful for patients with SCD).

Measure		Controls Mean (SD)	Patients Mean (SD)	p
Cold Detection (missing: 1 SCD)		29.2 (1.5)	29.1 (1.8)	0.695
Warm Detection (missing: 1 SCD)		34.5 (1.2)	35.2 (1.9)	<0.001
Cold Pain Threshold (missing: 2 SCD)		26.3 (5.0)	25.6 (5.6)	0.229
Cold Pain Intensity		2.0 (1.2)	2.5 (1.6)	<0.001
Heat Pain Threshold (missing: 2 SCD)		37.8 (3.6)	38.5 (4.0)	0.145
Heat Pain Intensity		2.1 (1.3)	2.6 (1.7)	0.005
	Filament Size	%	%
Mechanical Pain Threshold (missing: 2 SCD)	0.6	10%	17%	0.004
	1.4	15%	25%
	4.0	31%	22%
	6.0	7%	8%
	10.0	10%	8%
	26.0	8%	8%
	60.0	19%	12%
		Mean (SD)	Mean (SD)
Mechanical Pain Intensity (missing: 2 SCD)		0.7 (0.6)	0.9 (1.1)	0.007

Key: SCD = sickle cell disease; Controls = healthy, pain-free controls; Patients = patients with SCD

Prevalence of sensitization among controls and patients

We first compared the prevalence of sensitization on the reference test site between the controls and the patients (Table 4). For each patient, we tested two painful sites and one non-painful site; the reference site. The reference test site was the site at which the patients did not have current pain. For controls, we always used the anterior forearm site as the reference test site. For patients, this was not always possible because of preexisting pain at the anterior forearm, but an anterior forearm site (left or right) was the first choice. It turned out that we were able to use the anterior forearm site the vast majority (85%) of the time, followed by posterior forearm (10%) and upper arm lateral (5%). Compared with controls, the percentage of patients showing sensitization to at least one type of stimulus at this reference site was almost twice as high (23% patients vs 12% controls) and the difference was statistically significant (p=.02) (Table 4). This finding held across all sites (Table 4) and when we used only data from the anterior forearm as the reference site (25% patients vs 12% controls; p=.01).

Table 4.

Prevalence of sensitization among healthy control subjects and patients with sickle cell disease.

QST Test Site	Controls (N=124)	Patients (N=184)	p value
Reference site	12%	23%	.02
Across all sites	16%	32%	.004

Key: reference site was usually the anterior forearm, 2 patients lacked sufficient data to classify as sensitized

We also compared the groups of controls and patients on whether there was any sensitization across the three sites tested (Figure 2). Again, as shown in Table 4, the prevalence among patients (32%) was twice as high as the controls (16%), and the difference was statistically significant (p=.004).

Figure 2.

Prevalence of sensitivity for cold, heat, and mechanical pain thresholds across three test sites: Healthy controls (N=124) compared with patients with sickle cell disease (SCD, N=186)

Prevalence of sensitization among groups with various demographics

The prevalence of sensitization (overall) for controls and patients versus different demographic variables is shown in Table 5. We observed that college education was associated with lower prevalence among patients. Given the possible confounding of age and education level for younger subjects included in our study, we conducted a subgroup analysis of association between education and sensitization among subjects 25 years or older. The findings were substantially similar with no significant association among controls (p=.20) and significant association among patients (p=.007). Outcomes of binary logistic regression (Table 6) showed that there was a significant difference between controls and patients when sex and age were controlled; prevalence of sensitization was higher for patients.

Table 5.

Prevalence of sensitization (all sites) among different demographic groups of controls (N=124) and patients with sickle cell disease (N=184).

Variable	Category	Controls		Patients
		Prevalence	p values	Prevalence	p values
Sex	Female	18%	.63	37%	.08
	Male	14%		24%
Age	40+ years	20%	.33	38%	.14
	18–39 years	12%		28%
Education	High School	25%	.06	35%	.002
	Some College	6%		36%
	College+	13%		0%
Income	<$10,000	17%	.31	38%	.26
	$10,000-$30,000	24%		28%
	>$30,000	11%		24%

Table 6.

Logistic regression of sensitization prevalence.

QST Test Site	Predictor	Estimate	Std Error	t	p
Reference site	SCD (ref=controls)	0.825	0.335	2.467	.01
	Sex (ref=Female)	−0.355	0.313	−1.136	.26
	Age	0.031	0.012	2.494	.01
Across all sites	SCD (ref=controls)	0.887	0.297	2.988	.003
	Sex (ref=Female)	−0.469	0.278	−1.685	.09
	Age	0.021	0.011	1.910	.06

Key: reference site was usually the anterior forearm; SCD = sickle cell disease (N=184); Controls = healthy pain-free controls (N=124)

For patients, we also examined the association between sensitization and self-reported number of pain crisis episodes in the past 12 months. Thirty-nine percent of those with 4 or more pain episodes had sensitization versus 26% of those with 1–3 pain episodes and 18% of those with no pain episode (p=.04).

Types of sensitization among patients with SCD

For patients, we further applied our decision tree to classify them into four groups: normal (no sensitization at any site), peripheral sensitization, mixed sensitization, and central sensitization. The distribution of these 4 categories among patients as well as among patients by different age and sex subgroups appears in Table 7. We observed that among patients with sensitization, most exhibited some elements of central or mixed sensitization. Focusing on these two types of sensitization, we found that 79% of those with central sensitization exhibited sensitization in more than 1 QST modality/location versus 32% of those with mixed sensitization (p=.001). There was a statistically significant association between the sensitization category and the sex/age group combination (p<.01).

Table 7.

Sensitization classification among patients with sickle cell disease (*sensitized but not clear which category due to missing data).

Sensitization Classification	All Patients (N=184)	Patients Aged 18–39 years		Patients Aged 40+ years
		Female	Male	Female	Male
Normal	68%	65%	83%	60%	64%
Peripheral	1%	0%	0%	0%	4%
Mixed	12%	22%	0%	10%	11%
Central	18%	13%	17%	30%	14%
Unknown*	1%	0%	0%	0%	7%

Discussion

In a fairly large sample of patients with pain and SCD and age-, sex- and race-matched healthy controls, we used well-validated tools for thermal and mechanical QST and found that compared to controls, patients reported significantly higher sensitivity to pain threshold at the reference site as well as across all sites tested. Across all sex and age groups, the prevalence of sensitivity to pain threshold among patients who had pain related to SCD that was at least 3 on a 0–10 scale within the previous 12 months was consistently higher than controls at the reference site and across all sites. Controlling for age and sex, the prevalence of sensitization (allodynia or hyperalgesia to standard thermal and mechanical stimuli) in these patients remained significantly higher than that of the controls. Among patients, higher education levels were associated with lower overall prevalence of sensitization, and those with more frequent pain crisis episodes had sensitization in higher proportions than those with less frequent pain crisis episodes. Finally, our findings show that among these patients with pain and SCD, one third exhibited some element of sensitivity to pain threshold.

Our finding that compared to controls, patients reported a significantly higher sensitivity to pain threshold at the reference site as well as across all the sites tested was in the hypothesized direction. Using a much larger sample than in previous studies,^{7–9, 12} this finding supports the general conclusion of our previous study where patients with SCD reported allodynia and hyperalgesia to both thermal and mechanical stimuli,¹² an indication that sickle cell pain may have a neuropathic component. The inconsistency in the proportion of the sample with sensitization in our initial pilot study¹² and our current study may be a reflection in a difference in methods. In the pilot study, we conducted trial repetitions at the site without sufficient control of probe overlap, which may have contributed to wind-up. In the current study, however, there was no overlap in the test sites for in any of the trials. Our current findings are consistent with those of other researchers who found that compared to African American controls, patients with SCD exhibited higher prevalence of central sensitization.^7–9 Campbell and colleagues⁸ reported on 38 SCD patients (out of a larger sample of 83) with low (17; 20%) and high (21; 25%) central sensitization and found that the patients with high central sensitization were more likely to also report greater clinical pain than patients with low central sensitization. It is unknown, however, if the 32% prevalence of sensitization in our study is comparable to Campbell et al.⁸ since eligibility criteria and methods of classifying sensitization differed between the two studies. We specifically studied patients who had chronic pain to identify the prevalence of allodynia and hyperalgesia in that SCD population, which means that the actual prevalence rate for all patients with SCD could be lower than 32%.

The methodological difference in classification of sensitization is noteworthy. Campbell et al. conducted a battery of QST measurements including HPTh, HPTo (Heat Pain tolerance), PPTh (Pressure Pain Threshold), thermal and mechanical temporal summation, as well as hot water hand immersion. They then categorized SCD patients into low central sensitization, high central sensitization, and others by summarizing only the outcomes of thermal and mechanical temporal summation tests (the other QST measurements such as HPTh, HPTo etc. were not used in categorization). This focus is to elucidate wind-up and descending disinhibition mechanisms as a component of central sensitization.³⁹ In contrast, we utilized a set of measurements including HPTh, CPTh, MPTh, as well as HP, CP, and MP intensity. This focus is on peripheral sensitization and central sensitization related to facilitation of ascending nociceptive signals.³⁹ Specifically, we identified patients with allodynia or hyperalgesia as indicators of neuropathic pain, especially with emphasis on mechanical stimuli with the von Frey filaments, which exert much less pressure than the device used in Campbell’s study. The relative merits of the two approaches, however, remain to be answered by future research since sensitization is a complex phenomenon with a variety of ascending and descending mechanisms involving the peripheral neurons, dorsal root ganglia, presynaptic and postsynaptic spinal neurons, ascending and descending neurons and glial.^{3, 10, 23} Sorting out the specific mechanisms via QST in clinical populations such as SCD is challenging at this time. Our attempt to classify patients based on their responses to thermal and mechanical QST at two painful sites and one non-painful site is innovative but our early-stage findings warrant additional research for validation and to establish relevance regarding clinical pain assessment and management.

In view of the central sensitization in the sample with SCD, an important question that SCD researchers and clinicians must grapple with is whether prescribing opioids represents the optimal first line treatment for sickle cell pain among patients experiencing allodynia or hyperalgesia. We think not. The use of opioid analgesics as a mainstay therapy for controlling SCD pain was based on the notion that its etiology was only episodic, driven by vaso-occlusion and somatic or visceral tissue damage.^{2, 4} Evidence from the current study and previous work^{8, 9, 7, 12} on the contribution of sensitization to chronic pain in some adults with SCD suggests the need for a paradigm shift in the treatment of chronic pain in SCD. Medications that target pain mechanisms associated with allodynia and hyperalgesia, such as antidepressants, anticonvulsants, and other non-opioid analgesics should be considered in concert with opioid analgesics for optimal pain control in patients with SCD and indications of neuropathic pain, as is recommended for other patients with neuropathic pain.¹⁴

Few interventional studies have been published focused on the effects of these adjuvant therapies on the chronic pain of SCD that presents with indication of neuropathic components to the pain experience.⁵ In one study, the safety, maximum tolerated dose, and promising pain relief was shown for a CaMKIIα inhibitor, trifluoperazine, when used in adults with SCD who selected at least 4 pain descriptors that are thought to be indicative of neuropathic pain.²⁸ In another small study, pregabalin showed sustained reduction in pain over time within the treatment group, which suggests its promising effects for SCD pain.³⁴ Additional research is needed to study therapies for neuropathic pain since there is a growing body of preclinical and clinical research suggesting that neuropathic pain is likely to be part of the SCD pain experience for some patients.

Our QST findings showed that across all sex and age groups, compared with the controls, the prevalence of sensitivity to pain threshold among patient participants was consistently higher in both painful and the reference sites. Unlike controls, patients have had repeated ischemic tissue damage that could lead to sensitization. Further, the finding of no significant sex effects strengthens the interpretation of our results because the sensitization that patients reported could not be explained by the sex variable, which is in contrast with evidence from some researchers²⁶ but consistent with others.²⁴

We found a larger than expected prevalence of sensitization among controls. Our finding of 12% to 16% sensitization among controls is not unreasonable since 10% of community-based healthy participants reported signs and symptoms of neuropathic pain.⁴⁷ Although the sample of controls was relatively large for QST studies, replication and better understanding of the sensitization among the controls is warranted.

Although our findings are important, they should be interpreted in light of the study’s limitations. First, we did not evaluate whether patients with SCD have other nerve injuries that could contribute to their sensitivity to pain threshold. Consequently, we could not determine the mechanisms of the neuropathic pain and treatments that could be effective for it in patients with SCD. Given this limitation, the major implication of our findings is the need to design other prospective studies to determine the mechanisms contributing to the sensitization findings.

Second, we did not control for medications, including both analgesic medications such as use of opioids or hydroxyurea, which may influence the results. It will be important for future studies to determine the impact of medications on QST findings in SCD.

Third, from a large sample of African American controls and patients, both groups had QST responses closer to the adaptation temperature for both CPTh and HPTh than other samples reported in the literature. As well the vast majority of the sample reported pain from the ≤10 g von Frey filament stimuli, which convey light touch. One possible reason for these findings is that they followed the instructions we gave them to respond when they first felt pain. Supporting this explanation is the low average intensity they assigned to the pain from the stimuli (2.1 to 2.6 for HPTh, 2.0 to 2.5 for CPTh, and less than 1 for mechanical) and is much less than others found (average 5 for HPTh and 3.2 for CPTh).²⁵

Fourth, it is possible that our predefined thresholds (cutoffs) underestimated sensitivity. An ideal approach would be to examine different potential cutoffs and determine which cutoff has the best sensitivity and specificity for identifying sensitivity. However, there is no gold standard available against which to validate different cutoffs for sensitivity. The distribution of the individual measurements is highly skewed and non-normal, attributed to the heavy tail on the side away from the neutral point of 32 and ceiling/floor effect approaching the neutral point. As a consequence, mean and standard deviation, or even its robust version median and median absolute deviation are not good representations of the distribution. Therefore, the conservative nature of these cutoffs and the possibility that we are underestimating sensitization should be considered a study limitation despite our large samples for both controls and patients with SCD.

In conclusion, we present strong evidence supportive of a higher than normal prevalence of sensory sensitization in patients with SCD who have a history of pain (intensity of 3 or higher within the previous 12 months and >0 pain on at least one half of the days during the previous 3 months). Compared to the healthy controls, the evidence indicates higher prevalence of sensitivity to pain threshold in all sites tested across all sex and age groups among patients with SCD that remained significant controlling for sex and age. Future studies are needed to understand the mechanisms of this sensitization and find the optimal clinical diagnostic methods and treatments for its control among patients with SCD. In the meantime, findings indicate that allodynia and hyperalgesia can be part of the SCD pain experience and should be considered as therapies are selected for treating SCD pain.

Perspective.

Compared with matched healthy controls, quantitative sensory testing in adults with pain and sickle cell disease (SCD) demonstrates higher prevalence of sensitization, including central sensitization. The findings of allodynia and hyperalgesia may indicate neuropathic pain and could contribute to a paradigm shift in assessment and treatment of SCD pain.

Highlights.

Compared with controls, about twice as many adults with SCD showed sensitization.

Of those with sensitization, more than half exhibited some central sensitization.

Sensitization was associated with frequent pain crisis episodes among SCD patients.

Persistent allodynia and hyperalgesia can be part of the SCD pain experience.

Neuropathic pain, if part of SCD pain, should be appropriately assessed and treated.