The Definition, Assessment, and Prevalence of (Human Assumed) Central Sensitisation in Patients with Chronic Low Back Pain: A Systematic Review

Central sensitisation is assumed to be one of the underlying mechanisms for chronic low back pain. Because central sensitisation is not directly assessable in humans, the term ‘human assumed central sensitisation’ (HACS) is suggested. The objectives were to investigate what definitions for HACS have been used, to evaluate the methods to assess HACS, to assess the validity of those methods, and to estimate the prevalence of HACS. Database search resulted in 34 included studies. Forty different definition references were used to define HACS. This review uncovered twenty quantitative methods to assess HACS, including four questionnaires and sixteen quantitative sensory testing measures. The prevalence of HACS in patients with chronic low back pain was estimated in three studies. The current systematic review highlights that multiple definitions, assessment methods, and prevalence estimates are stated in the literature regarding HACS in patients with chronic low back pain. Most of the assessment methods of HACS are not validated but have been tested for reliability and repeatability. Given the lack of a gold standard to assess HACS, an initial grading system is proposed to standardize clinical and research assessments of HACS in patients with a chronic low back.


Introduction
The worldwide prevalence of chronic low back pain (CLBP) ranges between 2-25% [1][2][3]. According to the 2017 global burden of disease study low back pain is the leading cause of years lived with disability [4]. Recent evidence suggests that assessments of pain might represent underlying pain mechanisms and possibly help to identify patients at risk of poor response to different treatments for low back pain [5,6]. Currently, there are three mechanistic descriptors to describe patients' pain: (a) nociceptive pain [7]; (b) neuropathic pain [7]; and (c) nociplastic pain [8]. Nociplastic pain has been suggested to cover a subset of patients with CLBP with widespread hyperalgesia, facilitated temporal summation of pain, and/or impaired conditioned pain modulation, without clear evidence of actual or threatened tissue damage, that are less likely to respond positively to standard pain treatments [9,10]. Before the term nociplastic pain was introduced, terms used included centralised pain [11,12], hyperresponsiveness [13], central hypersensitivity [14], and central sensitisation (CS) [11]. via e-mail from the study authors. When the study consisted of patients with various chronic pain condition(s), the data for patients with (solely) CLBP was requested by e-mail from the authors.

Study Descriptives
The study descriptives included country, number of included patients, and patient characteristics (age, sex, and Body Mass Index (BMI)). The studies were divided into three groups: patients with CLBP only, patients with CLBP in combination with other pain condition(s) (CLBP+), and healthy volunteers (only when a comparison was made between patients with CLBP and healthy volunteers).

Definitions of Human Assumed Central Sensitisation
The definitions used for HACS and the source references for these definitions were collected.

Assessment of Human Assumed Central Sensitisation
Methods for the assessment of HACS were included when they were used for the assessment of HACS in the methods section of the included study. Studies that proposed new assessment methods or a translation of a current assessment method for HACS were also eligible for appraisal if this was stated in the introduction. Potential new assessment methods for HACS mentioned in the discussion section of the studies were excluded because the initial use, as described in the method section, of the assessment method was different in the study. Therefore the assessment method was not examined or validated as an assessment method for HACS in that study.

Clinimetrics of Human Assumed Central Sensitisation Assessment Methods
Clinimetrics outcomes, including sensitivity, specificity, accuracy, internal consistency, the area under the curve, likelihood ratio, predictive values, and test-retest reliability, were retrieved from the included studies. Associations with other assessment methods in these studies, independently if these other assessment methods were HACS-related, were extracted. In this review, a correlation coefficient of less than 0.29 was deemed low, from 0.3 to 0.49 was considered moderate, from 0.5 to 0.69 substantial, from 0.70 to 0.99 very high, and 1.0 perfect [39,40].

Prevalence of Human Assumed Central Sensitisation
The prevalence of HACS was retrieved when explicitly reported in eligible studies. Additionally, the CSI part A scores were extracted to assess the prevalence of HACS using a cut-off score of 40, according to previous suggestions [25].

Data Synthesis
All extracted data were synthesised to provide an overview of the information for each study aim, study characteristics, and heterogeneity in the study findings. To calculate the prevalence of HACS in patients with CLBP, the CSI scores, presented in the articles or provided by the authors, were used. Both the calculated prevalence and those presented in the included articles were also used to calculate an overall prevalence. When available, the prevalence for CLBP only and CLBP in combination with other pain condition(s) were calculated separately.

Results
A total of 12,764 articles were identified. Checking the reference lists of the included studies resulted in 280 more articles. After the removal of duplicates, 8772 articles remained, and 507 titles were selected to be screened on abstract. 153 articles were included in the full-text screening. This resulted in the inclusion of 34 studies in this review, as shown in the PRISMA Flow Diagram (Figure 1).
A total of 12,764 articles were identified. Checking the reference lists of the i studies resulted in 280 more articles. After the removal of duplicates, 8772 art mained, and 507 titles were selected to be screened on abstract. 153 articles were i in the full-text screening. This resulted in the inclusion of 34 studies in this rev shown in the PRISMA Flow Diagram (Figure 1).   Table 1 shows descriptive data from the included studies. These data represent the number of patients with CLBP only, CLBP+, and healthy controls. In 16 out of 34 studies, it was not specified whether patients either had CLBP only or CLBP+, in these studies 'CLBP' was stated because no distinction could be made from the data provided. For all groups, the number of patients, age, sex, and BMI (when applicable) are presented (Table 1).

Risk of Bias
The results of the QUADAS-2 are shown in Table 2 and Figure 2. Of the 34 included studies, 33 did not describe a reference standard for HACS. Therefore, this was assessed as a high risk of bias for risk of bias and applicability. For the risk of bias, it was not possible to assess "flow and timing" [25,. Only 1 study used a reference standard, however, this study did not compare its results to an index test for HACS [71]. The reference standard they used was a clinical judgement which, as they noted, may be used as an appropriate alternative reference standard in the absence of a 'diagnostic' gold standard [71,74,75]. An index test was not compared to a reference standard for HACS in any of the included studies, hence "flow and timing" was noted as not applicable.

Definition of Human Assumed Central Sensitisation
In the studies included in this review, there were various definitions postulated for HACS. In 3 of the studies, no definition was provided [61,72,73]. The references used for the definitions were retrieved to find the similarities and differences between the different

Definition of Human Assumed Central Sensitisation
In the studies included in this review, there were various definitions postulated for HACS. In 3 of the studies, no definition was provided [61,72,73]. The references used for the definitions were retrieved to find the similarities and differences between the different definitions. In total, 40 definition references were used to define HACS. The number of definition references used per study varied between 1 and 9. Most definitions referenced that used by Woolf (2011) [18]: "an amplification of neural signalling within the central nervous system that elicits pain hypersensitivity" (n = 17). The second most frequently used definition reference was from IASP (2008) [7]: "Increased responsiveness of nociceptive neurons in the central nervous system to their normal or subthreshold afferent input" (n = 6). The third most frequently used definition reference was from Nijs et al. (2010) [76]: "an augmentation of responsiveness of central neurons to input from unimodal and polymodal receptors" (n = 5). All other definition references were used three times or less, as shown in Table 3. Table 3. Definition used to describe human assumed central sensitisation and the reported prevalence of human assumed central sensitisation.  "Central sensitisation is an amplified state of neural signalling in the central nervous system (CNS) that is implicated in the pathogenesis of several chronic conditions that primarily involve pain and complex, multisymptom illnesses. When in the sensitized state, the CNS amplifies the sensory processing of the peripheral inputs so that the experience of the individual no longer accurately reflects the information provided by peripheral inputs. This state has been described as an increase in signal gain in which low-level sensory inputs are amplified into stronger signals, or as a decrease in signal inhibition processes, or both."   In the abstract: "Central sensitisation has been proposed as a common pathophysiological mechanism to explain related syndromes for which no specific organic cause can be found." In the introduction: "Central sensitisation, which involves an abnormal and intense enhancement of pain by mechanisms in the central nervous system, maybe the common link between these disorders."

Assessment of Human Assumed Central Sensitisation
Four questionnaires and 16 QST methods to assess HACS were observed (Table 4). In Table 5, correlations greater than or equal to 0.5 between HACS assessment outcome and other assessment methods from the studies, are shown (Table 5). Some assessment methods have described correlations greater than or equal to 0.5 as well as below 0.5 in other studies. This counts for the Oswestry Disability Index, the physical and mental component of the short form-36, NRS pain intensity, McGill Pain Questionnaire, Pain Catastrophizing Scale, sensory profiles: sensory seeking and low registration, and State-Trait Anxiety Inventory. All correlations are presented in Table S2.   Scores from the WPI and SS are combined to determine the presence and severity of CS.
Not reported As part of the 2011 FM survey [110,111] Cutoff scores of ≥ 12 with a combination of either WPI score ≥ seven and SS score ≥ five or WPI score 3 to 6 and SS score ≥ 9 distinguish those with CS as FM positive. Higher total scores indicate a greater degree of CS.
The Michigan Body Map (MBM) [107] was used in one study [62] to assess HACS. The MBM is a patient-reported outcome measure to assess body areas where chronic pain is experienced and to quantify the degree of widespread body pain when assessing pain features. The MBM, was used to determine the extent to which pain had spread throughout the body as a possible result of HACS [62]. If only one body area was affected, the likelihood that HACS was present was minimal, but the more body areas affected, the greater the chance that HACS was present.
The short-form McGill Pain Questionnaire revised version (SF-MPQ-2) [109] was used for the assessment of HACS in one study [62]. The SF-MPQ-2) has a mechanism-based approach to assess neuropathic pain [108,109], and was used in this study as a proxy for HACS [62]. The authors, however, did not specify how the SF-MPQ-2 was used to assess HACS.
The Widespread Pain Index (WPI) [110,112] was used in one study to determine the extent that pain had spread throughout the body as a possible result of HACS [43]. With an increasing number of affected body areas, the possibility of the presence of HACS also increases. In this study, the 2011 FM survey, a diagnostic tool for FM, consisted of both the WPI and symptom severity [110,111].

QST Measures
Pressure pain threshold (PPT) was assessed using a handheld algometer at up to four different anatomical locations [43,44,60,61,64,72,73]. The PPT was conducted using a push force gauge whereby pressure applied over a particular muscle or tissue is increased until the patient describes the pressure as becoming painful. In these studies, HACS was indicated when PPTs were lower in patients with CLBP compared to healthy controls. PPT was assessed in the originally affected area or on distal test areas, for example, the opposite side of the body or another body part. The anatomical assessment locations differed by study, and the number of locations assessed differed with a range of 1 to 4 locations (Table 4). In three studies, patients with CLBP were compared with healthy volunteers [43,72,73].
Temporal summation (TS) or wind-up ratio was used four times [54,64,72,73] to estimate HACS. As wind-up ratio and TS refer to the same phenomenon, the term TS is in this review used for both. Three different methods were used to assess TS: (1) TS with heat pain [54], (2) TS with predetermined PPT [64], and (3) TS with pinprick [72,73]. To calculate TS, the mean pain ratings of the series were divided by the mean pain ratings of single stimuli. TS was considered positive when the pain intensity increased within the test series.
Conditioned pain modulation (CPM) [43,60,61] was used in three out of thirty-four studies to assess HACS. CPM is a human proxy assessment to evaluate the balance between the descending inhibitory and facilitatory pathway activity [43,60,61,113,114]. For CPM, the effect of applying a second painful stimulus in addition to the first was assessed; the first painful stimulus is termed the test stimulus, and the second painful stimulus was referred to as the conditioning stimulus. A comparison between test stimuli can be made before, after [60], and during the conditioning stimulus [43,60]. Two different conditioning stimuli were utilized by the studies in this review. These were: (1) a pressure cuff [43], and (2) ice water [60,61]. PPT was used in all the studies as a test stimulus in different locations; thumbnail [43], lower back [43], upper leg [60], and 2nd toe [61].
Other QST measurements. In two studies [72,73], in addition to the previously mentioned QST measures, extra QST measures were obtained (Table 4), including (A) mechanical detection threshold, (B) Mechanical pain threshold, (C) Mechanical pain sensitivity, and dynamic mechanical allodynia, and (D) Vibration detection threshold. According to these studies, the QST measures assess all relevant aspects of the somatosensory system, including large and small fibre functions and signs of central sensitisation.
Although authors claimed to address HACS in six studies [45,52,53,63,66,71], they did not describe assessment methods that could estimate HACS in the method section of their study.

Estimation of the Prevalence of Human Assumed Central Sensitisation
The prevalence of HACS in patients with CLBP was described in three studies, as shown in Table 3 (Table 3). The CSI scores resulted in a prevalence of 71.1% (n = 128) [46]. The presence of tactile allodynia resulted in a prevalence of 60.8% (n = 74) in patients with CLBP+, a prevalence of 13.3% (n = 15) in patients with CLBP only [52] and PPT combined with TS resulted in a prevalence of 18.3% (n = 104)) [64]. Based on these 3 studies (n = 321), the mean prevalence of HACS in patients with CLBP was 48.9%.

Discussion
The aim of this systematic review was to highlight outcome assessments used for HACS in patients with CLBP. As a gold standard for HACS assessment does not exist, it was important to gauge the various terminologies applied to HACS as well as assessment methods in humans used in literature. A variety of definitions and assessment methods were observed. Evidence supporting the preference for a particular assessment method or a combination of assessment methods was not found. The QUADAS-2 showed for risk of bias mostly unclear or low risk of bias scores. The reference standard was scored in almost all studies as a high risk of bias because only once a reference standard was used. There was no comparison between the index test and the reference standard, therefore all flow and timing were not applicable. Applicability concerns mainly showed a low risk of bias for patient selection and index test, and the reference standard was mostly a high risk of bias, because no reference standard was used.

Definition of Human Assumed Central Sensitisation
Across all the studies concerning HACS, a wide variety of definitions for CS are described. At this moment it can't be said which definition of CS in patients is to be used. Most definitions as described in the included studies in this review are more explanatory rather than really defining central sensitisation with all the associated issues. This is because central sensitisation is more than the outcome of just one assessment. Moreover, central sensitisation might also reflect an adaptive primary mechanism or may occur as a secondary response following an (acute) injury [80]. Finally, (nociplastic) pain is not the same as central sensitisation and central sensitisation might not be the main cause of chronic pain [115].
The definition proposed by the International Association for the Study of Pain [100] cannot directly assessable in humans, as direct recordings from central neurons based on electrophysiological methods are not possible yet. Therefore the term central sensitisation should not be used as such in both human research and clinical practice [24].
Although an HACS-related term was used in many articles included in this review, these terms were only stated in the introduction and discussion sections of six articles [45,47,51,56,63,66]. This gives the impression that the authors used HACS as an explanatory model for the findings from their assessment methods, but lacked a clear explanation of how they considered and defined HACS. Twenty studies, did not use an HACS-related term in the results section [25,42,45,47,48,50,51,53,54,56,58,60,61,63,[65][66][67][68]72,73], and therefore did not provide additional value for the assessment of HACS, the clinimetrics of assessment methods and/or the prevalence of HACS. The variety of definitions and ambiguous terminology used in the literature made the assessment of HACS even more unclear because a lack of unequivocal definitions and terms makes the assessment of HACS more confusing and incorrect. In 2019 a more comprehensive definition for CS was presented: "Hyperexcitability of the central nervous system" [23]. This definition assumed that the central, not peripheral, nervous system is the main contributor to HACS. However, this cannot be established in most cases of HACS, because altered sensation can be caused peripherally as well as centrally. It is therefore clear that a unified definition for HACS was not utilized by the included studies.

Methods for Assessing Human Assumed Central Sensitisation
The manifestation of HACS might refer to the possible presence of CS as a neural mechanism in humans. HACS can present itself in many symptoms and/or manifestations such as altered CPM, decreased pain thresholds, increased temporal summation, and the presence of widespread pain) but can, until today, not directly be demonstrated in humans [24]. In three studies, patients were appointed to groups with and without HACS based on CSI or QST assessment outcomes [46,52,64]. In all other studies, methods were used to assess HACS, but no conclusion about the presence or absence of HACS could be drawn. The CSI is a questionnaire that is often used, also in the included studies of this review, and can be a quick tool for identifying whether symptoms may be associated with HACS or syndrome/disorder/disease related to HACS. High CSI scores may suggest HACS or suggest the need to perform additional tests to assess syndrome/disorder/disease related to HACS, after which appropriate treatment can be initiated [116]. However, because of the lack of a reference standard to date, no scientific evidence to support the value of the CSI as an indicator for HACS can be found [64,117]. Hence, the interpretation of the CSI data should be made cautiously. None of the included studies provided evidence, by relating the CSI to other HACS assessments, that could establish a significant role for the CSI in the assessment of HACS. Moreover, the cut-off score is established at 40 out of 100 [84]. It should be noted that there is a lack of cut-off scores for QST resulting from highly variable assessment methods between research groups and a lack of high-quality, large sample normative data. Different assessment methods can be compared to each other, such as questionnaires with physical assessment methods. In this review, only one significant substantial negative/inverse correlation was found (CSI and lifting capacity) [42]. With respect to the various methods assessing HACS, this review did not find clear associations between assessment methods for HACS or other non-HACS assessments methods in the studies. It should be noted that some studies discuss the presence of HACS based on assessment methods used in the study without initially describing this in the method section. These were excluded from this analysis because validation or examination of these assessments for HACS was not possible.

Clinimetrics of Human Assumed Central Sensitisation Assessment Methods
Clinimetrics concerning only reproducibility and consistency were calculated in 7 studies. Sensitivity and specificity were reported only once [64]. The clinimetric outcomes reported in most included studies concerned internal consistency and test-retest reliability which were high for the CSI. Sensitivity and specificity of the assessment methods are required to address the validity of an assessment method; in the included studies they were, however, not described likely due to the aforementioned lack of diagnostic gold standard.
A new cut-off value of 20 (0-100) for the CSI for patients with HACS was presented in the study of Mibu et al. [64] with a sensitivity of 73.7% and a specificity of 37.7%, not supporting this new cut-off as being sufficient to discriminate for HACS. Neblett et al. [84] reported a sensitivity of 81% and a specificity of 75% for the CSI, but included patients with chronic pain and not only patients with CLBP. In the same study, a non-patient group (undergraduate students not being treated for chronic pain) was compared with a group of patients with one or more syndrome/disorder/disease related to HACS in order to identify the cut-off value [84]. Nebletts' study does not improve the potential discriminability of HACS with the CSI, as it was not tested in a group of patients with chronic pain only. Mibu et al. [64] based the presence of HACS on the PPT tertile data and a positive TS value. Using the lowest tertile of the PPT score can give a distorted picture because the cut-off value may change and vary based on the assessed population.

The Estimated Prevalence of Human Assumed Central Sensitisation
The prevalence estimates 43.2% of HACS in patients with CLBP in the literature [41,42,46,47,[49][50][51][55][56][57][58][59]65,[67][68][69][70] should be interpreted as being based on the assumption that specific symptoms and signs reflect mechanisms that are observed in animal experiments and could explain pain symptoms for a subset of patients experiencing CLBP. Therefore, we introduced the term HACS and have estimated its prevalence in patients with CLBP as derived from proxy assessments.
Few studies in this review estimated the prevalence of HACS, which varied between 13% and 78% [46,52,64]. These prevalence estimates should be interpreted with caution as calculations were based on several different assessment methods, i.e., CSI scores, tactile allodynia, and PPT scores combined with TS scores. The CSI is so far the only assessment method with an established cut-off value, however, the CSI has been suggested to show a stronger association to pain severity and psychopathology than with QST measures of HACS [118]. All studies are inherently limited due to a lack of a gold standard and, thus, limited available measurement properties. Therefore, it should be noted that the discriminability and, consequently, the prevalences are debatable.
To obtain a better understanding of the effects of co-existent chronic pain condition(s) on the prevalence of HACS in patients with CLBP, the prevalence of HACS between patients with CLBP only and patients with CLBP in combination with other pain condition(s) was compared. These prevalence estimates were based mostly on the CSI, however, the prevalence based on tactile allodynia showed similar scores. Due to the low number of studies (n = 8) and the low number of patients (CLBP only n = 708 and CLBP + n = 752) used to estimate prevalence in this review, these findings should be interpreted with caution.

Limitations
This review focused on patients with CLBP, but in some studies, patients were also included with CLBP in combination with other pain condition(s). Due to the complex nature of chronic pain, and often difficulties in diagnosing it, pain condition(s) were not always well-documented in the included studies. It is possible that HACS is presented more in patients with neuropathic pain, this might result in less generalisability of the data. But the presence of neuropathic pain is not well documented in all included studies, therefore, a separate analysis was not possible. Therefore, forming the analysis groups for studies with 'CLBP only' or "CLBP in combination other pain condition(s)" could only be completed in 16 studies [41,42,45,46,49,50,53,54,60,61,[64][65][66]69,72,73]. The possible presence of multiple pain conditions in patients may have influenced the results. Similarly, HACS was considered as a static phenomenon with prevalence estimates based on a single assessment. Although reliability was high for HACS assessment methods when reported, recent work has suggested that psychological symptoms associated with the CSI [105], as well as PPT and TS [119], may fluctuate with pain severity. For now, it is still unclear if HACS is a state or a trait. Both human and animal studies would suggest the presence of CS changes over time depending on pain presence and pain progression [19].
It was part of the original setup of this study to do a meta-analysis. However, because of the vast number of varying assessments methods found in all included studies, while each of these methods was applied just once or a few times it was impossible to do this analysis.
The risk of bias assessment based on the QUADAS-2 was challenging to use in combination with the selected studies. The lack of a gold standard for defining and estimating HACS, the reference standard, and flow and timing domains hinders adequate judgement. However, this was the case for all studies and did not affect the comparison of the risk of bias assessment between papers. The Newcastle-Ottawa-Scale [120] was considered as a possibility to be used in the assessment of the risk of bias, but none of the studies were cohort or case-control studies. No other suitable risk-of-bias assessment method for diagnostic accuracy studies was found.

Gold Standard
To date, CS cannot according to the IASP taxonomy be demonstrated in humans, and HACS as a proxy lacks a uniform definition and is an unambiguous way of assessment. The lack of a gold standard in patients with CLBP has led to the application of a large variety of methods to assess HACS as shown in this review. Lack of clarity on how to assess HACS is illustrated by the finding that QST is referred to as a gold standard in one of the studies [121]. However, in other studies this statement is refuted [48,55,59,60,68,71,80,122,123]. Until CS can be demonstrated in humans a gold standard cannot be established. Researchers are still searching for an acceptable proxy to establish HACS [122]; examples are the CSI [48,55,59,60,123], exclusion of neuropathic pain and the differential classification of nociceptive versus HACS pain [81,124], QST measures [60,68] and clinical judgement [64]. For patients with CLBP, one study was included in this review where the assessment methods were compared with an expert's opinion [71]. In this study, based on an expert's opinion, patients with CLBP were classified founded on posture, movement, and neurological assessment, and a clinical interview. Here the authors did an attempt to assess HACS by excluding another mechanistic pain that did not apply.
Another major challenge and limitation is to differentiate HACS from peripheral sensitisation. When comparing assessments on various body locations, for example, distant from the primary pain location, it should be noted that an assessment such as PPT might reflect peripheral sensitisation when assessed in only one or two locations and more assessment locations are recommended. As shown in this review, there are many studies with different definitions and different assessment methods, leading to inconclusive results, which shows a need for a uniform way of assessing HACS. The question arises if and how the expert's opinion can be used as a reliable assessment to be considered a proxy to assess HACS. To meaningfully contribute, this expert opinion should be based on a uniform definition for HACS, a uniform way of HACS presentation, and a uniform way of assessing HACS.

A Grading System for Human Assumed Central Sensitisation
The scientific discussion about the assessment of neuropathic pain is much more advanced than for the assessment of HACS. In neuropathic pain, a diagnostic gold standard for assessment is also lacking, and although many questions are still unanswered, there seems to be some more understanding of this clinical condition compared to HACS. A grading system provides a more uniform way to grade the probability of neuropathic pain in patients in daily clinical practice as well as in research. This grading system for neuropathic pain, firstly described by Treede et al. [30] and updated by Finnerup et al. [29], is widely accepted and serves as a basis to propose an idea to assess HACS more uniformly: a grading system for HACS (Figure 3).  Similar to the neuropathic pain grading system, the idea behind a grading system for HACS is to incorporate the various definitions of HACS, the available (clinical) tests, and the (validated) assessment methods to become a more directive assessment of HACS in daily clinical practice as well as in (clinical) research. The grading system for HACS as presented here should be further discussed, elaborated on, further developed, and validated for clinical and scientific use in future studies.
At the first level, patients with CLBP are graded based on the patient's history. When patients' pain is only partly explained by anatomical, morphological and neurological lesions, and/or diseases or fully unexplained, there is a possibility of HACS. This is because HACS can be present in patients with nociplastic pain, but also in patients with nociceptive or neuropathic pain. If this criterion is fulfilled, the first level of certainty-"possible" HACS is reached.
The second level is based on the indicator tests (see Table 6). With an increasing number of positive indicator tests, as described in this review, the description of "probable" HACS may be relevant. When one or no indicator tests are fulfilled, the patient has "unconfirmed" HACS. Among the described tests to assess HACS in this review, there are Similar to the neuropathic pain grading system, the idea behind a grading system for HACS is to incorporate the various definitions of HACS, the available (clinical) tests, and the (validated) assessment methods to become a more directive assessment of HACS in daily clinical practice as well as in (clinical) research. The grading system for HACS as presented here should be further discussed, elaborated on, further developed, and validated for clinical and scientific use in future studies.
At the first level, patients with CLBP are graded based on the patient's history. When patients' pain is only partly explained by anatomical, morphological and neurological lesions, and/or diseases or fully unexplained, there is a possibility of HACS. This is because HACS can be present in patients with nociplastic pain, but also in patients with nociceptive or neuropathic pain. If this criterion is fulfilled, the first level of certainty-"possible" HACS is reached.
The second level is based on the indicator tests (see Table 6). With an increasing number of positive indicator tests, as described in this review, the description of "probable" HACS may be relevant. When one or no indicator tests are fulfilled, the patient has "unconfirmed" HACS. Among the described tests to assess HACS in this review, there are assessments that can be used as an indicator test for HACS in patients with CLBP. Candidates are the CSI, PPT, CPM, and TS. It should be noted that there is currently only a cut-off value for the CSI established based on previously performed validation studies. For the other indicator tests, the validity should be established to be able to distinguish between the presence or absence of HACS. Future studies should also be aimed at assessing the validity of multi-modal test batteries to assess HACS. The third level should grade patients with CLBP based on diagnostic tests or confirmatory tests as "definite" HACS. To assess the level of "definite" HACS, more scientific research is required because there are currently no diagnostic methods to assess HACS which are validated against a gold standard. This third level might be added in the future when confirmative assessment methods are available. However, promising insights with respect to HACS are already shown in brain imaging such as fMRI [22] or blood serums such as BDMR [27,28].

Limitations Initial Grading System
HACS cannot and should not be seen as a 'diagnosis' in humans. Assessment of HACS via the grading system will raise the possibility of the presence of HACS which is based on the different manifestations and symptoms related to HACS. In addition, raising the possibility of HACS is not based on a single, specific, assessment but is based on following a process using various assessment methods. Future research is needed to further validate, improve, and expand the initial grading system for HACS. This is to create a uniform process in the assessment of HACS. This will lead to a more comparable (clinical) assessment of patients, which will benefit daily clinical practice as well as research on this topic. As an example of such studies, previously performed forward and backward translations between animal studies and human studies regarding the assessment of CS showed that the cuff-algometer test for the assessment of CPM is similar to what is assessed in animals (rats) and in humans [125]. For the suggested indicator tests (PPT, WUR, CPM) clinically relevant values should be established. Other assessment methods that are less frequently used to assess HACS should be evaluated for possible expansion of the indicator tests. Furthermore, confirmatory tests should be studied to confirm the presence of HACS. With the addition of confirmatory tests, the indicator tests could be improved. However, it is still needed to find objective assessment methods for HACS and a gold standard method that can be used in daily clinical practice.

Conclusions
In this systematic review, various definitions of HACS and assessment methods are discussed with a specific focus on CLBP. It may be concluded that the definitions of HACS that apply to chronic pain in humans, in general, show no uniformity in consensus. Moreover, there is no standardised way to assess HACS in patients with CLBP based on current assessments that use various proxies. Given the fact that CS cannot, according to the definition, be measured directly in humans, further research is needed to develop reliable proxies. A gold standard to diagnose HACS is lacking, undermining the validity of the actually available indicator tests for HACS in patients with CLBP, but an initial grading system for HACS is proposed.

Data Availability Statement:
The authors of the original papers can request all data and related metadata underlying this review's findings. Data can also be requested from the first author of this systematic review, but data will only be provided with the original data authors consent.