Table 1.
Selected baseline characteristics and prognosis across centers with different preferences for primary decompressive craniectomy.
| Treatment preference (observed primary DC rates per centre)a |
|||||
|---|---|---|---|---|---|
| Quartile 1 (6–12%) | Quartile 2 (12–19%) | Quartile 3 (19–26%) | Quartile 4 (26–67%) | SMD | |
| n | 53 | 48 | 51 | 48 | |
| Age (median [IQR]) | 63 [56, 69] | 56 [43, 66] | 56 [38, 68] | 53 [34, 64] | 0.26 |
| Sex | 0.29 | ||||
| Female | 12 (23) | 15 (31) | 23 (45) | 9 (19) | |
| Male | 41 (77) | 33 (69) | 28 (55) | 39 (81) | |
| ASAPS (%) | 0.44 | ||||
| Healthy | 28 (53) | 17 (35) | 28 (55) | 23 (48) | |
| Mild systemic disease | 21 (40) | 16 (33) | 14 (27) | 15 (31) | |
| Severe systemic disease | 4 (8) | 11 (23) | 7 (14) | 7 (15) | |
| Threat to life | 0 (0) | 0 (0) | 1 (2) | 0 (0) | |
| Unknown | 0 (0) | 4 (8) | 1 (2) | 3 (6) | |
| Hypoxia (%)b | 0.49 | ||||
| No | 41 (77) | 42 (88) | 44 (86) | 37 (77) | |
| Definite | 1 (2) | 1 (2) | 5 (10) | 6 (12) | |
| Suspect | 2 (4) | 4 (8) | 1 (2) | 2 (4) | |
| Unknown | 9 (17) | 1 (2) | 1 (2) | 3 (6) | |
| Hypotension (%)c | 0.44 | ||||
| No | 41 (77) | 46 (96) | 46 (90) | 39 (81) | |
| Definite | 2 (4) | 1 (2) | 3 (6) | 2 (4) | |
| Suspect | 1 (2) | 0 (0) | 1 (2) | 3 (6) | |
| Unknown | 9 (17) | 1 (2) | 1 (2) | 4 (8) | |
| Any major extracranial injury (%)d | 23 (43) | 15 (31) | 28 (55) | 25 (52) | 0.27 |
| GCS baseline (median [IQR]) | 9 [4, 13] | 7 [3, 11] | 5 [3, 9] | 6 [3, 11] | 0.30 |
| GCS motor baseline (median [IQR]) | 5 [2, 6] | 4 [1, 5] | 1 [1, 4] | 2 [1, 5] | 0.43 |
| Pupils (%) | 0.32 | ||||
| Both reacting | 36 (68) | 35 (73) | 32 (63) | 28 (58) | |
| One reacting | 3 (6) | 7 (15) | 9 (18) | 10 (21) | |
| Both unreacting | 14 (26) | 6 (12) | 10 (20) | 10 (21) | |
| Any neuroworsening before surgery (%) | 15 (28) | 11 (23) | 13 (25) | 16 (33) | 0.12 |
| Total volume of ASDH (cm3, median [IQR]) | 58 [31, 97] | 70 [40, 114] | 70 [32, 103] | 50 [18, 79] | 0.24 |
| CT large ASDH (%)e | 35 (66) | 37 (77) | 42 (82) | 31 (65) | 0.25 |
| CT midline shift (%)f | 42 (79) | 38 (79) | 48 (94) | 44 (92) | 0.29 |
| CT contusion (%) | 0.40 | ||||
| No | 19 (36) | 25 (52) | 21 (41) | 23 (48) | |
| Small | 24 (45) | 19 (40) | 23 (45) | 13 (27) | |
| Large | 10 (19) | 3 (6) | 6 (12) | 11 (23) | |
| Unknown | 0 (0) | 1 (2) | 1 (2) | 1 (2) | |
| CT subarachnoid haemorrhage (%) | 0.37 | ||||
| No | 18 (34) | 21 (44) | 12 (24) | 21 (44) | |
| Basal | 5 (9) | 2 (4) | 6 (12) | 2 (4) | |
| Cortical | 22 (42) | 15 (31) | 27 (53) | 16 (33) | |
| Basal and cortical | 8 (15) | 10 (21) | 6 (12) | 9 (19) | |
| CT basal cisterns absent/compressed (%) | 20 (38) | 19 (40) | 25 (49) | 21 (44) | 0.13 |
| Mean predicted 6-month unfavourable outcome (GOS score ≤3, %, median [IQR])g | 74 [52, 86] | 73 [53, 87] | 80 [67, 91] | 69 [51, 84] | 0.22 |
| Centre characteristics | |||||
| Academic hospital (vs. non- academic, %) | 0 (0) | 0 (0) | 0 (0) | 14 (29) | 0.45 |
| Number of beds (median [IQR]) | 655 [600, 850] | 1083 [1018, 1148] | 1170 [936, 1292] | 780 [652, 831] | 0.80 |
| Residency program neurosurgery (%) | 53 (100) | 48 (100) | 51 (100) | 48 (100) | <0.01 |
| Level I trauma centre designation (%) | 42 (100) | 48 (100) | 51 (100) | 35 (100) | <0.01 |
| Urban location (vs. suburban and rural location, %) | 53 (100) | 48 (100) | 51 (100) | 48 (100) | <0.01 |
| Neurosurgeon staffing (FTE, median [IQR]) | 11 [8, 19] | 11 [10, 12] | 10 [6, 12] | 8 [8, 11] | 0.50 |
| Number of surgeries for ASDH in 2013 (median [IQR]) | 28 [10, 30] | 18 [16, 20] | 62 [20, 102] | 25 [22, 25] | 0.82 |
| Number of surgeries for contusion in 2013 (median [IQR]) | 7 [5, 8] | 10 [7, 14] | 15 [4, 236] | 10 [8, 14] | 0.37 |
| Low threshold policy for primary DC in ASDH (%)h | 0 (0) | 0 (0) | 0 (0) | 24 (50) | 0.71 |
Abbreviation: AIS, Abbreviated Injury Scale; ASAPS, American Society of Anesthesiologists classification system; ASDH, acute subdural hematoma; DC, decompressive craniectomy; GCS, Glasgow Coma Scale; GOS, Glasgow Outcome Scale (5-point); IQR, interquartile range; IV, instrumental variable; SMD, standardised mean difference.
Treatment preference as defined by the case-mix adjusted probability of undergoing primary DC (as opposed to craniotomy) based on the observed primary DC rates per centre. This corresponds to the IV status and presented in quartiles of the range of adjusted regional primary DC rates. The first category is less aggressive than the second and the second is less aggressive than the third and so forth. Importantly, the IV analysis used adjusted primary DC rates as continuous preference, the quartiles are presented for purposes of interpretability of baseline comparability.
Second insult during the pre-hospital or ER phase, defined as PaO2 < 8 kPa (60 mmHg)/SaO2 < 90%. ‘Suspected’ was scored if the patient did not have documented hypoxia by PaO2 or SaO2, but there was a clinical suspicion, as evidenced by for example cyanosis, apnoea or respiratory distress.
Second insult during the pre-hospital or ER phase, defined as systolic BP < 90 mmHg. ‘Suspected’ was scored if the patient did not have a documented blood pressure, but was reported to be in shock or have an absent brachial pulse (not related to injury of the extremity).
AIS ≥3.
Large is defined qualitatively by the treating neurosurgeon and corresponded to a size larger than 25 cm³.
Midline shift present is classified as being more than 5 mm.
TBI severity as summarised in predicted unfavorable outcome, proportion with a Glasgow Outcome Scale ≤3, based on CRASH-CT variables age, GCS score, pupillary reactivity to light, major extracranial injury, and CT characteristics (midline shift >5 mm, traumatic subarachnoid hemorrhage, and obliteration of the basal cisterns).
Before patient inclusion in CENTER-TBI, treatment policies per centre were captured by provider profile surveys, including the policy towards primary DC. The resulting threshold for primary DC is dichotomised based on this distinction: Yes', primary DC routinely/pre-emptively versus ‘No’, no primary DC routinely/pre-emptively.