The study was conducted in three distinct phases, which involved professionals from six Intensive Neurorehabilitation Units, qualified by the Italian National Health System for the care of post-acute patients with sABI. The participating centers are part of the Department for rehabilitation of patients with sABI at Fondazione Don Gnocchi ONLUS (Italy).¹

In the first phase, clinicians with high expertise in the management of the post-acute phase of sABI developed the final version of FDG-CCS. The FDG-CCS was elaborated based on a previous longitudinal study that aimed to record the presence of the most frequent clinical complications in patients with sABI and their impact to patients’ outcomes (6). In the Estraneo et al.’s study (6), clinical complications were grouped in 10 categories. The first six categories, modeled on the CIRS (18), included the clinical complications affecting the main systems of the body (i.e., endocrine-metabolic, cardio-vascular, musculo-skeletal-cutaneous, gastro-intestinal, genito-urinary tract, respiratory). Four additional categories included clinical conditions that more frequently and specifically occur in patients with sABI (i.e., neurosurgical complications, epilepsy/myoclonus, heterotopic ossification, and paroxysmal sympathetic hyperactivity). The severity of clinical complications was classified as moderate or severe on the basis of the intensity of the treatment protocol required (in categories 1–8); severity of heterotopic ossification (in category 9) was related to the number of affected joints; paroxysmal sympathetic hyperactivity (category 10) was only classified as present or absent (see Table 1).

The FDG-CCS differed from the previous version of the checklist because it included 13 categories of clinical complications and introduced a numeric rating system. The two items “endocrine-metabolic disorders” and “musculo-skeletal-cutaneous complications” of the previous checklist were split in four separate categories (i.e., endocrine disorders, metabolic disturbances, musculo-skeletal and cutaneous complications), to assess their occurrence more specifically. Indeed, available data show that metabolic alterations and endocrine disturbances can specifically and separately influence patients’ clinical evolution (19, 20), whereas cutaneous (e.g., pressure sores) and musculo-skeletal complications are both particularly frequent in patients with sABI but imply tailored treatment and differentially impact on outcomes (21). Moreover, a new category “sepsis,” i.e., “life-threatening organ dysfunction caused by a dysregulated host response to infection” (22) was added, as patients with acute neurologic injury are usually at higher risk for developing in hospital sepsis, due to frequent need for ventilatory support and intravenous therapy via central line catheters that increases the exposition to repeated interactions with multi-resistant pathogens in the lower respiratory trait and in the bloodstream (23). Sepsis has been found to increase mortality in the acute phase of patients with traumatic brain injury (24) and to be the most common cause of in-hospital deaths in several countries (23). Sepsis was diagnosed according to the quick SOFA score criteria [i.e., respiratory rate of 22/min or greater, altered mentation, or systolic blood pressure of 100 mm Hg or less (22)], regardless of the source of infection. Presence of infection of each system/apparatus is reported in their specific categories (3–9), but, for instance, in case of a severe urinary tract infection leading to sepsis, a score is assigned both for the urinary category and for sepsis category.

As for the numeric system for rating clinical complexity of patients with sABI, the FDG-CCS classifies severity of medical complications included in groups 1–9 on a 4-point ordinal scale, with score 0 meaning lack of complication, whereas scores 1–3 are assigned based on the requirement and intensity of the therapeutic intervention: 1 = no treatment required (mild complication), 2 = required pharmacological or non-pharmacological treatment (e.g., dressing for pressure sores) not associated with intensive clinical monitoring (moderate complication), and 3 = required urgent and/or continuous or sub-continuous treatment and clinical monitoring or surgical interventions (severe complication). The 4-point score (i.e., 0–3) is assigned based on frequency and duration of the episodes for categories 10 (epilepsy/myoclonus) and 11 (paroxysmal sympathetic hyperactivity), and as a function of the number of affected joints for category 12 (heterotopic ossifications). Lastly, the category 13 (sepsis) is rated 0 if absent or 1 if present. In case of occurrence of several clinical complications within the same category, only the most severe is scored.

The FDG-CCS total score ranges from 0 (i.e., no complications) to 37 (presence of all clinical complications with the highest severity). A detailed description of the FDG-CCS and the comparison with original categorization are reported in Supplementary material and Table 1.

In the second phase, a single online training session was performed with professionals from the six neurorehabilitation units, including several professionals who did not participate in the development of the FDG-CCS. This phase was dedicated to present administration and scoring procedures for the Italian version of the FDG-CCS and the CoCoS. Subsequently, all the participating professionals were required to administer and score FDG-CCS on 3 patients with sABI, whose detailed medical reports and paraclinical evaluations (e.g., laboratory tests, radiological exams) performed in a 15-day hospital stay were provided by the study coordinator. Professionals’ scores were analyzed to detect potential inconsistencies among examiners raters and any difficulty in the administration protocol were discussed. At the end of the training phase an inter-rater agreement was calculated for each of the 3 cases. Inter-rater agreement was poor for the score of a clinical complication in the gastrointestinal category (constipation), as the treatment administered (i.e., evacuation enema) was classified as moderate in half of the examiners and urgent in the other half. In the evaluated patients, the correct score was 2 (moderate severity), as it was sufficient to administer the treatment once, and without need of intensive monitoring. Another issue was raised about the possible difficulty to classify some metabolic alterations as metabolic or endocrine. For instance, hypo- or hyper-glycemia could be classified in metabolic category and as a marker of an endocrine disorder (i.e., diabetes mellitus). However, in case of coexistence of metabolic alterations and endocrine disorders, the FDG-CCS protocol recommended scoring only the clinical sign (e.g., hypo- or hyper-glycemia) if the related endocrine disorder (e.g., diabetes mellitus) has not been diagnosed, otherwise scoring only the endocrine disorder (see Supplementary material).

In the third phase of the study, three professionals (A, B, and C) for each participating center independently administered and scored the FDG-CCS (examiners A and B) and the CoCoS (examiner C) in a sample of sABI inpatients, to assess the inter-rater agreement and concurrent validity of the FDG-CCS.

Within each of the six participating intensive neurorehabilitation units, three professionals who were potential or actual caretakers of sABI patients were identified as examiners A, B and C in the validation study. Although physicians have specific expertise on clinical complication management, we also involved non-physician professionals for two main reasons: first, we aimed at proposing a well operationalized tool, with simple and clear scoring criteria to be applied in usual rehabilitation settings; second, in rehabilitation settings non-physician professionals are usually informed about patient’s clinical complications, treatment, rehabilitation program and outcomes during the periodic team meetings, so that they are in the position of contributing to assessment of patients’ clinical complexity. All recruited examiners (n = 18) had prior experience in the care of patients with sABI; they worked as physicians (n = 13), physiotherapists (n = 2), speech therapists (n = 2) or nurses (n = 1). Each participating center enrolled the first 7 consecutively admitted patients with sABI in a period of 6 months (May–October 2024) and who met the following inclusion criteria: (i) age ≥ 18 years; (ii) clinical diagnosis of sABI (i.e., coma >24 h with Glasgow Coma Scale ≤8 after traumatic, vascular, or anoxic brain injury); (iii) time post-injury ≥28 days; (iv) written informed consent by the patient or by her/his legal representative when the patient could not give the consent him/herself.

Since we aimed at validating a novel clinical tool, we had no reference in the literature for estimating the expected effect size. For this reason, we computed the sample size focusing on the correlation between the FDG-CCS and the CoCoS, assessed by Spearman’s rank correlation coefficient (ρ). As we hypothesized the presence of a correlation (i.e., discarding the null hypothesis of ρ ≠ 0), although of small-moderate entity (i.e., ρ = 0.50), between the scales, sample size calculation was done assuming as null hypothesis (H0) the absence of correlation (ρ = 0) and as alternative hypothesis (H1) the existence of correlation between the two scales (ρ ≠ 0). The sample was estimated by the G*Power software (25), using a t-test for correlation and using a two-tailed test and assuming a correlation between the scales of ρ = 0.5, a type I error of 0.05 and a power of 90%. On these bases, 42 patients were recruited, 20% of them to cope with possible dropouts.

At study entry (i.e., enrolment), the following patients’ data were collected: i. demographic (age, sex); ii. medical history (time post-injury, etiology, CIRS comorbidity and severity scores prior to the brain injury); iii. Clinical features (diagnosis based on standardized clinical criteria of VS/UWS, MCS minus, MCS plus, or emergence from DoC) (5, 26); consciousness level measured by the best total score (out of at least three in a week) of the Italian version of the Coma Recovery Scale-Revised (CRS-R) (27); functional level assessed by means of the Glasgow Outcome Scale-Extended score (GOS-E) (28); level of cognitive functioning score (LCF) (29); disability level assessed by the Disability Rating Scale total score (DRS) (30).

On the 15th day from enrolment, two independent examiners (examiners A and B) administered the FDG-CCS taking into account the previous 2 weeks as observation time. This procedure was implemented in order to level out the probability of occurrence of clinical complications among patients enrolled from the different centers, which could be influenced by time post-injury and by duration of hospitalization. For the purposes of the study, the examiners scored both novel incident complications and acute relapses or worsening of pre-existing medical comorbidities if they required moderate (i.e., pharmacological or non-pharmacological treatment but not intensive clinical monitoring) or intensive (i.e., urgent with a continuous or sub-continuous clinical monitoring) treatment during the two-week observation period. Concurrently, a third examiner (examiner C), blind to the score of the FDG-CCS of A and B examiners, administered the CoCoS on the same observation period. The 3 examiners scored the items of the FDG-CCS and of the CoCoS only based on information documented in the medical records, by laboratory tests and/or instrumental examinations or by direct clinical observation.

The CoCoS is a measurement tool designed specifically to assess comorbidities in patients with pDoC on recovery (17). The CoCoS consists of 24 categories addressing the frequency of various comorbidities common in this subset of patients. The categories include respiratory and urinary tract infections, non-infectious respiratory disorders, structural heart diseases, rhythm disorders without structural heart diseases, arterial hypertension, diabetes mellitus, dysautonomia, peripheral artery or venous diseases, with a specific separated category for those of supra-aortic trunks, hepatobiliary and gastrointestinal disorders, seizures, hydrocephalus, fractures and joint diseases, anemia, presence of life support devices, pressure ulcers, malignancies, malnutrition, renal diseases, and previous disability. For each category, scoring is based on the presence/absence of the comorbidity and its severity. Severity is scored based on presence of symptoms, the need for treatment and the response to it. The severity due to the need of life support devices is scored considering the number of necessary devices (tracheostomy tube, nasogastric tube, percutaneous endoscopic gastrostomy, urinary catheter, central venous catheter). The cumulative comorbidity burden is denoted by the final summative score ranging from 0 (no comorbidities) to 72 (presence of all comorbidities at maximum of their severity). Based on the final cumulative score, patients are stratified in 3 severity groups using the following cut-off scores: (0: no comorbidities; range 1–24: presence of mild comorbidities; range 25–48: moderate comorbidities; range 49–72: severe comorbidities).

The investigators in charge of data collection also took care of the data entry. Patients’ data were collected and entered in a pseudonymised form into a centralized, password-protected, electronic database on REDCap (Research Electronic Data Capture), following as much as possible a forced-choice format. REDCap complies with the Health Insurance Portability and Accountability Act in compliance with security measures and therefore only contains data that has already been pseudonymised. Indeed, the personal data of each patient were de-identified and replaced by an alphanumeric ID code in an ‘association key’ document accessible exclusively to the enrolling center in a password-protected Excel file and then entered into REDCap. Moreover, REDCap ensures blindness between two examiners, as each examiner is not allowed to see the other’s assessment.

Demographic and clinical data were submitted to Shapiro–Wilk tests to investigate the normality of the distributions. As these preliminary analyses showed that not all continuous variables followed a normal distribution, except for age, the descriptive data of the sample on admission to the study were expressed in terms of mean ± standard deviation for age, median (inter-quartile range, IQR) for all other continuous variables, and as frequencies for categorical variables. Parametric or non-parametric analyses were performed accordingly. In particular, demographic and clinical variables were compared between conscious and pDoC patients, and between VS/UWS and MCS patients, by means of t-test, the Mann–Whitney U test or χ²-test, as appropriate.

The FDG-CCS total score of both examiner A and B was compared between conscious and pDoC patients, and between VS/UWS and MCS patients by means of Mann–Whitney U tests.

For the assessment of inter-rater agreement in the administration of the FDG-CCS, Cohen’s K was calculated on the scores of the individual sub-scales and the intra-class coefficient (ICC) was calculated on the total scores of the FDG-CCS, administered by the two examiners (A and B). For Cohen’s K, the values of 0.4 or less can be considered poor, values between 0.4 and 0.6 moderate, values between 0.6 and 0.8 good, whereas values greater than 0.8 suggest excellent inter-observer agreement (31). For ICC, values less than 0.5 are indicative of poor reliability, values between 0.5 and 0.75 indicate moderate reliability, values between 0.75 and 0.9 indicate good reliability, and values greater than 0.90 indicate excellent reliability (32).

For the test of concurrent validity between the FDG-CCS (administered by both examiners A and B) and the CoCoS (administered by examiner C), Spearman’s rank correlation coefficient was calculated on the total scores obtained by the patients on the two scales.

The results were considered statistically significant if p < 0.05. All analyses were performed using IBM SPSS v.25 (IBM Corp., Armonk, NY, United States).

Forty-two patients with sABI were enrolled in the study, including 22 patients with pDoC and 20 fully conscious patients at study entry. Conscious patients and patients with pDoC did not differ in terms of age, sex, etiology, CIRS severity and comorbidity scores, but significantly differed in terms of time post-injury, clinical diagnosis, CRS-R, LCF, GOS-E, and DRS scores. Table 2 presents the demographic and clinical data of the participants, categorized by clinical diagnosis of consciousness. Within the pDoC group, patients in VS/UWS did not differ from those in MCS in terms of demographic or anamnestic characteristics (all p > 0.05), but the two diagnostic sub-groups differ significantly in the CRS-R [U = 2.5; p > 0.005; VS/UWS median = 6 (IQR = 2) vs. MCS = 12 (4)], LCF [U = 10.5; p = 0.001; VS/UWS = 2 (0) vs. MCS = 3 (9)], GOS-E [U = 28.5; p = 0.043; VS/UWS = 2 (0) vs. MCS = 3 (1)], and DRS [U = 25.0; p = 0.025; VS/UWS = 24 (0) vs. MCS = 22 (3)] scores.

The frequency and severity distribution of clinical complications evaluated by the FDG-CCS, as assessed by both examiners A and B, is shown in Figure 1. The median total score for examiners A was 3 (IQR = 5), whereas for examiner B was 3 (IQR = 4). The FDG-CCS total score did not differ between conscious and pDoC patients, nor between VS/UWS and MCS, for both examiners A and B (all p > 0.05).

Metabolic (Examiner A = 33%; Examiner B = 43%), gastro-intestinal (A = 31%; B = 26%), cardio-vascular (A = 26%; B = 29%), respiratory (A = 21%; B = 21%), and musculo-skeletal disorders (A = 19%; B = 14%) were the most frequent complications; sepsis was reported in 2% of the cases.

The only complications that were rated as severe by both examiners A and B were respiratory and paroxysmal sympathetic hyperactivity, in 5% (n = 2/42) of the cases. On average, the complications were rated as mild in 4.3% (mean percentage between examiners) of the cases, and moderate in the 9.9% of the cases. The complications that were rated as moderate more frequently were metabolic (27%) and gastro-intestinal (27%).

Metabolic complications were more frequent in conscious (Examiner A = 45%; Examiner B = 65%) than in pDoC patients (A = 22.7%; B = 22.7%; A-χ² = 11.617; A-p = 0.003; B-χ² = 10.526; B-p = 0.005). No significant differences in frequency or severity of any FDG-CCS category have been observed between VS/UWS and MCS (all p > 0.05).

Inter-rater agreement for the total score of the FDG-CCS was good. Inter-rater agreement for single FDG-CCS categories was excellent for cardio-vascular, musculo-skeletal, gastro-intestinal, epilepsy/myoclonus, heterotopic ossification, paroxysmal sympathetic hyperactivity, and sepsis; good for cutaneous and respiratory complications; moderate for endocrine, genito-urinary tract, and neurosurgical complications; poor for metabolic complications (see Table 3 for detailed values).

Physicians (n = 13) who scored the FDG-CCS items as examiners A and B showed lower inter-rater agreement with respect to non-physicians (n = 4; i.e., 1 physiotherapist, 2 speech-therapists, 1 nurse). In particular, non-physicians provided an excellent inter-rater agreement for endocrine, cardio-vascular, musculo-skeletal, cutaneous, gastro-intestinal, respiratory, neurosurgical, epilepsy/myoclonus, heterotopic ossification, paroxysmal sympathetic hyperactivity, and sepsis (all K = 1.00; p < 0.005); good for genito-urinary tract (K = 0.725; p < 0.005); moderate for metabolic complications (K = 0.481; p < 0.005); and an excellent reliability for the total score (ICC = 0.989; p < 0.005).

Conversely, physicians provided an excellent inter-rater agreement for cardio-vascular (K = 0.873; p < 0.005), gastro-intestinal (K = 0.859; p < 0.005), epilepsy/myoclonus, heterotopic ossification, paroxysmal sympathetic hyperactivity, and sepsis (all K = 1.00; p < 0.005); good for musculo-skeletal (K = 0.797; p < 0.005), cutaneous (K = 0.650; p < 0.005), and respiratory complications (K = 0.605; p < 0.005); moderate for genito-urinary tract (K = 0.520; p = 0.002) and neurosurgical complications (K = 0.481; p < 0.005); poor for metabolic (K = 0.251; p = 0.070) and endocrine complications (K = 0.385; p = 0.001); and a good reliability for the total score (ICC = 0.773; p < 0.005).

As for concurrent validity, the median total score of the CoCoS was 8 (IQR = 8). The FDG-CCS total score correlated significantly with the total score of the CoCoS, for both the examiner A (ρ = 0.356; p = 0.01) and B (ρ = 0.317; p = 0.02).