Table 1.
Summary of studies evaluating the effect of glucocorticosteroids for acute pancreatitis on clinical score, duration of hospitalization, C-reactive protein, and histopathologic score
| Study | Level of evidence, study design and methodological quality (for details see Table 2) | Species | Number of subjects in groups | Acute pancreatitis type | Treatment: Type of GC, dose, duration, administration route and timing | Other treatment (GC group + model controls) | Data collection and timing, post induction unless otherwise stated | Outcome: Clinical score | Outcome: CRP | Outcome: Duration of hospitalisation | Outcome: Mortality | Outcome: Pancreas histopathology |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Okanishi et al. [57] |
LOE III Non-RCT Moderate estimated risk of bias, Moderate number of well characterised dogs in each group |
Dogs |
Model controls: 20 GC group: 45 |
Spontaneous AP | Prednisolone 1 mg/kg/day SC, from diagnosis to discharge | IV fluids, maropitant, famotidine, enrofloxacin, fentanyl, multivitamin solution | Daily clinical score evaluation in hospital, CRP, mortality. Follow up 1 month after diagnosis | Days until clinical score ≤ 2/3 was significantly lower in GC group (P < 0.001) (median 4 vs 7 days) | Days until CRP had reached < 2 mg/dL was significantly lower in GC group (P < 0.001) (median 4 vs 8 days) | Significantly shorter in GC group (P = 0.002) (median: 5 vs 8 days) | 1-month survival was significantly higher in GC group (P = 0.005) (88.7% vs 57.9%) | NA |
| Studley and Schenk 1982 [39] Experiment 1 |
LOE II RCT, low estimated risk of bias, very small number of poorly characterised dogs in each group |
Dogs |
Model controls: 8 Group 1: 6 Group 2: 6 |
Induced AP: 10 mL bile injected into the pancreatic duct |
Hydrocortisone 2.86 mg/kg IV, then 1.43 mg/kg q8h Controls: No GC Group 1: GC 6 h post induction Group 2: GC 12 h post induction |
IV fluids: 5% dextrose in 0.2% saline 10 mL/kg SC q8h | Survival up to 72 h | NA | NA | NA |
Survival time increased in treatment groups controls: 18.8 h Group 1: 40.2 h (P < 0.05) Group 2: 48.3 h (P < 0.03) No difference between treatment groups |
NA |
|
Studley and Schenk [39] Experiment 2 |
LOE II RCT, low estimated risk of bias, very small number of poorly characterised dogs in each group |
Dogs |
Model controls: 6 Group 1: 5 Group 2: 6 |
Induced AP: 10 mL bile injected into the pancreatic duct |
Hydrocortisone Controls: No GC Group 1:2.86 mg/kg IV, 6 h post induction, then 1.43 mg/kg q8h Group 2: Hydrocortisone 14.3 mg/kg 6 h post induction, then q8h |
Fluid replacement: 5% dextrose in 0.2% saline 10 mL/kg SC q8h | Survival up to 72 h | NA | NA | NA |
Survival time increased in treatment groups Controls: 27 h Group 1: 45.6 h (P < 0.05) Group 2: 41.6 h (P < 0.05) No difference between treatment groups |
NA |
| Attix et al. [58] |
LOE III Controlled trial, high estimated risk of bias, very small number of poorly characterised dogs in each group |
Dogs |
Model controls: 4 GC group: 4 |
Induced AP: Oleic acid 5 mL/kg injected into the main pancreatic duct | Dexamethasone 0.06 mg/kg IV q24h for 8 days, start 24 h post induction | IV fluids incl. glucose for 5 days postop. Procaine penicillin G 20.000 U/kg IM q12h, Kanamycin sulfate 5 mg/kg IM q8h, NPO for 8 days | Clinical evaluation 3 times daily, histopathology at day 8, technique and results not shown, only photos | No clinical difference | NA | NA | NA | No results/conclusions for difference in histopathology among groups reported |
| Stewart et al. [59] |
LOE III Controlled trial, high estimated risk of bias, small number of poorly characterised dogs in each group |
Dogs |
Model controls: 16 GC group: 25 |
Induced AP: A solution of trypsin and 4% sodium taurocholate injected into the main pancreatic duct | Cortisone acetate 100 mg IM, as soon as the abdomen could be closed, then 50 mg q12h for 4 days, then gradually withdrawn | None stated | Survival, pancreas histopathology when survival was assured (day 5–60), technique and results not shown | NA | NA | NA |
Survival: control 6%, GC group 48% |
GC group had decreased oedema and inflammation in the first few days, less apparent in animals autopsied after ≤ 5 days |
|
Imahori et al. [60] Study 1 |
LOE III Controlled trial, low estimated risk of bias, small number of poorly characterised dogs in each group |
Dogs |
Model controls: 14 GC group: 14 |
Induced AP: Autologous bile injected into the pancreatic duct | Hydrocortisone 2.86 mg/kg IV, 6 h post induction, then 1.34 mg/kg q8h | Fluid treatment 5% dextrose in 0.2% saline 10 mL/kg SC q8h | 72 h survival, histopathology at death or 72 h, score 0–3 (oedema, PMN leucocyte infiltration, vascular thrombosis, necrosis, haemorrhage) | NA | NA | NA | Significantly increased duration of survival in GC group (P < 0.01) | Acinar necrosis significantly decreased in GC group (P < 0.05) |
|
Imahori et al. [60] Study 2 |
LOE III Controlled trial, high estimated risk of bias, very small number of poorly characterized dogs |
Dogs |
Model controls: 6 GC group: 5 |
Induced AP: Autologous bile injected into the pancreatic duct | Hydrocortisone 14.3 mg/kg at 6, 14 and 22 h post induction | Fluid treatment 5% dextrose in 0.2% saline 10 mL/kg SC q8h | Ultrastructural changes to the pancreas at 24 h | NA | NA | NA | NA | Pancreatic structural integrin better preserved in the GC group |
| Kaplan et al. [65] |
LOE III Retrospective case control study, high estimated risk of bias, small number of fairly well characterised patients in each group |
Humans |
Model controls: 48 GC group: 15 GC |
Spontaneous Acute oedematous: 51 Acute haemorrhagic/necrotising: 12 |
Hydrocortisone 100-200 mg IV every 6-8 h for 2–3 days, then tapered | Nasogastric suction, anticholinergics, antibiotics, IV fluids | Clinical evaluation, mortality | Dramatic response reported | NA | NA |
Oedematous AP: all survived Haemorrhagic AP: Overall mortality 8/12 (67%) GC group 5/9 (55%) |
NA |
| Liu et al. [40] |
LOE II Randomised controlled trial, moderate estimated risk of bias, moderate number of poorly characterised patients in each group |
Humans |
Model controls: 53 GC group: 26 |
Spontaneous AP | Dexamethasone 20–30 mg then 30-60 mg/day, tapered over 3–5 days, Dextran 500 mL/day and Salvia miltiorrhiza 20–30 mL, treatment initiated at diagnosis | NPO, pancreatin secretion inhibition, antibiotics, fluids, parenteral nutrition | Mortality | NA | NA | NA | Reduced in GC group 11.54% vs 32.08% (P < 0.05) | NA |
| Eklund et al. [37] |
LOE III Retrospective case control-study, low estimated risk of bias, small number of well characterised patients in each group |
Humans |
Model controls: 11 GC group: 10 |
Spontaneous AP | Hydrocortisone 3–400 mg/day, started 4-11d after symptom onset | Ventilation, IV fluids antibiotics (cefuroxime mostly) | CRP up to 48 h, 30-day mortality | NA | CRP significantly lower in GC group at 48 h (P = 0.043), not at 24 h | NA |
30-day mortality GC group 30%, controls 36% |
NA |
| Wang et al. [66] |
LOE IV Case-series, high estimated risk of bias, moderate number of fairly well characterised patients in the group |
Humans |
Model controls: 0 (literature reference) GC group: 32 |
Spontaneous AP |
Dexamethasone 0.5–1 mg/kg q24h for3-5 days, starting 8 h to 4 days after onset of symptoms Dextran-40 500–1000 mL/day for 7 days |
Nasogastric tube decompression, oxygen, IV fluids, antibiotics (imipenem) | Clinical evaluation, mortality | 4-8 h after treatment start pain was relieved | NA | NA | GC group: 12.5% literature: 40% | NA |
| Wan et al. [38] |
LOE II RCT, low estimated risk of bias, moderate number of well characterised patients in each group |
Humans |
Model controls: 35 GC group: 35 |
Spontaneous AP | Dexamethasone IV 1 mg/kg q6h for 3 days | NPO, IV fluids, analgesia, proton pump inhibitors, parenteral nutrition, modified Dachengqi Decoction, plasma and albumin if necessary, 14d antimicrobials | Observation and follow up for 1 month | NA | NA |
Duration of hospitalisation significantly reduced in GC group (P = 0.03) 32.5 vs 40.2 days |
GC group: 8.6%, model controls: 14.3% (NS) | NA |
| Zhang et al. [41]* |
LOE II RCT, moderate estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 45 GC group: 45 |
Induced, sodium taurocholate 3.5% 1 mL/kg injected into bile-pancreatic duct | Dexamethasone 5 mg/kg IV single dose, 15 min post induction | None stated | Survival at 3,6,12 h, pancreas histopathology score 0–4 (oedema, acinar necrosis, inflammation, perivascular infiltrate, haemorrhage, fat necrosis) | NA | NA | NA | No difference in mortality | Pancreas histopathology score lower in GC group vs model controls at 3 and 6 h (P < 0.05), and at 12 h (P < 0.01) |
| Schulz et al. [42] |
LOE II RCT, low estimated risk of bias, small number of well characterised rats in each group |
Rats |
Model controls: 15 GC group: 15 |
Induced AP, sodium taurocholate 3.5% 2 mL/kg injected into the biliary-pancreatic duct | Methylprednisolone 25 mg/kg IV, 1 h post induction, followed by CRI 0.125 mg/kg | Ringers lactate 0.5 mL/h | Survival, histopathology at 20 h score 0–3 (oedema, inflammatory infiltration, haemorrhage, necrosis, fat necrosis) | NA | NA | NA | No difference in survival | No difference in histopathology |
|
Gloor et al. [61] Experiment 1 |
LOE III Controlled trial, low estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 32 GC group: 32 |
Induced AP, cerulein 5 µg/kg/h over 6 h IV (oedematous pancreatitis) | Hydrocortisone 10 mg/kg IV 10 min post induction | Fluid replacement | Histopathology at 1.5 h, 3 h, 6 h, 12 h; score 0–4 (oedema, inflammatory infiltration), 0–7 (necrosis, haemorrhage) | NA | NA | NA | NA | No difference in histopathology score, results not shown |
|
Gloor et al. [61] Experiment 2 |
LOE III Controlled trial, low estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 40 GC group: 40 |
Induced AP, sodium taurocholate 5% injected into the biliopancreatic duct (necrotising pancreatitis) | Hydrocortisone 10 mg/kg IV 10 min post induction | Fluid replacement | Histopathology at 1.5 h, 3 h, 6 h, 12 h; score 0–4 (oedema, inflammatory infiltration), 0–7 (necrosis, haemorrhage) | NA | NA | NA | NA | No difference in histopathology score, results not shown |
|
Gloor et al. [61] Experiment 3 |
LOE III Controlled trial, low estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 8 GC group: 21 |
Induced, sodium taurocholate 5% injected into the biliopancreatic duct (necrotising pancreatitis) | Hydrocortisone 10 mg/kg IV 10 min post induction | Fluid replacement | 72 h survival | NA | NA | NA | Reduced in GC group versus model controls (P = 0.01) | NA |
| Cosen-Binker et al. [43] |
LOE II RCT, moderate estimated risk of bias, small number of well characterised rats in each group |
Rats |
Model controls: 16 Group 1: 16 Group 2: 16 |
Induced AP, sodium taurocholate 7% injected into the biliopancreatic duct |
Hydrocortisone SC 4 h post induction Model controls: No GC Group 1: GC 4 mg/kg Group 2: GC 2 mg/kg |
None stated | Histopathology at 8 h post induction, score 0–4 (oedema, haemorrhage, leucocyte infiltration, acinar necrosis, fat necrosis) | NA | NA | NA | NA | Unclear description of results numerical and statistical differences were found |
| Cosen-Binker et al. [44] |
LOE II RCT, low estimated risk of bias, small number of well characterised rats in each group |
Rats |
Model controls: 16 Group 1: 16 Group 2: 16 Group 3: 16 Group 4: 15 |
Induced AP, sodium taurocholate 8% 1 mL injected into the biliopancreatic duct |
Hydrocortisone 6 mg/kg SC or Prednisolone 0.5 mg/kg SC 1 h or 4 h post induction Controls: No GC Group 1: HC 1 h Group 2: HC 4 h Group 3: Pred 1 h Group 4: Pred 4 h |
None stated | CRP, histopathology at 10 h, score 0–4 (oedema, haemorrhage, leukocyte infiltration, acinar necrosis, fat necrosis) | NA |
HC or pred 1 h post induction improvement of CRP (P < 0.05) HC or pred 4 h post induction no difference |
NA | NA | HC or pred 1 h post induction improvement of histopathology score (P < 0.05). HC or pred 4 h post induction no difference |
| Jha et al. [45] |
LOE II RCT, high estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 24 GC group: 24 |
Induced AP, sodium taurocholate 4% 1 mL/kg injected into the biliopancreatic duct | Dexamethasone 0.5 mg/kg IV after induction | None stated | Histopathology at 3 h, 6 h and 12 h, score 0–4 (oedema, acinar necrosis, haemorrhage, fat necrosis, inflammation, perivascular infiltration), TEM of pancreatic tissue | NA | NA | NA | NA |
Authors claim marked difference in histopathology and TEM pathologic changes between groups at 12 h Results not shown |
| Melo et al. [46] |
LOE III RCT, low estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 8 GC group: 8 |
Induced AP, l-arginine 2.5 g/kg IP twice with 1 h interval | Methylprednisolone 30 mg/kg PO 1 h after induction | None stated | Histopathology at 24 h, score 0–3 (oedema, leukocyte infiltration, haemorrhage, acinar vacuolization & necrosis) | NA | NA | NA | NA | Reduced histopathology score on all parameters in GC group compared to model controls (P < 0.05) |
| Ramudo et al. [62] |
LOE III Controlled trial, moderate estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 6 GC group: 6 |
Induced AP, bile-pancreatic duct obstruction (BPDO) | Dexamethasone 1 mg/kg IM 1 h post induction | Buprenorphine 0.2 mg/kg IM | Histopathology at 3 h and 12 h, score 0–3 (oedema, inflammatory cells, vacuolisation, necrosis) | NA | NA | NA | NA | Reduced oedema and leucocyte infiltration at 12 h in GC group compared to model controls (P < 0.05) |
| Ramudo et al. [47] |
LOE II RCT, low estimated risk of bias, unknown number of well characterised rats in each group |
Rats |
Model controls: not stated GC group: not stated |
Induced AP, sodium taurocholate 3.5% 0,1 mL injected into the biliopancreatic duct | Dexamethasone 1 mg/kg IM 1 h post induction | Buprenorphine 0.2 mg/kg IM | Histopathology at 3 h and 6 h, score 0–3 (oedema, inflammatory cells, vacuolisation, necrosis) | NA | NA | NA | NA | Reduced necrosis at 6 h in GC group compared to model controls (P < 0.05) |
| Ou et al. [48]** |
LOE II RCT, high estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 36 GC group: 36 |
Induced, sodium taurocholate 3.5% 0.1 mL/100 g injected into the biliopancreatic duct | Dexamethasone 0.5 mg/100 g IV 15 min post induction, then CRI 0.05 mg/100 g/h (dose only mentioned in Zhang et al. 2010) | None stated | Mortality rate, Pathological severity score at 3 h, 6 h, 12 h (not further explained), but results shown | NA | NA | NA | Lower in GC group compared to model controls at 12 h (P < 0.05) | Reduced histopathology severity score at 12 h in GC group compared to model controls (P < 0.05) |
| Zhao et al. [49] |
LOE II RCT, high estimated risk of bias, small number of well characterised rats in each group |
Rats |
Model controls: 15 GC group: 15 |
Induced AP, sodium taurocholate 5% injected into the biliopancreatic duct | Dexamethasone 5 mg/kg IV following induction | None stated | Histopathology, scoring technique not stated, results not shown | NA | NA | NA | NA | Author claims Oedema and necrosis reduced in GC group compared to model controls especially at 12 h |
| Foitzik et al. [50] |
LOE II RCT, moderate estimated risk of bias, small number of well characterised rats in each group |
Rats |
Model controls: 10 Group 1: 10 Group 2: 10 Group 3: 12 |
Induced AP: Glycodeoxycholic acid 10 mM injected into the biliopancreatic duct, then IV 5 µg/kg/h cerulein over 6 h |
Prednisolone started 6 h after induction, q8h for 24 h Ccontrols: No GC Group 1: GC 2 mg/kg/day Group 2: GC 10 mg/kg/day Group 3: GC 50 mg/kg/day |
IV fluids 6 mL/kg/h for 8 h after induction | Mortality, histopathology 6 h after last treatment, score 0–4 (oedema, acinar necrosis, haemorrhage, fat necrosis, inflammation, perivascular infiltrate) | NA | NA | NA | Mortality 20–40%, no difference among groups | Histopathology score for acinar necrosis: No difference among groups |
| Biradar and Veeresh [51] |
LOE II RCT, high estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls:6 GC group: 6 |
Induced: 2 Intraperitoneal injections of l-arginine 2.5 g/kg, 1 h apart | Methylprednisolone 30 mg/kg/day PO 1 h post induction | None stated | CRP histopathology 24 h after last L-arginine injection, score 0–3 (oedema, acinar cell degeneration, interstitial inflammation, haemorrhage), | NA | CRP results for GC group not show | NA | NA | Authors claim less oedema, Inflammation, acinar cell degeneration & necrosis in GC group, results not shown |
| Biradar and Veeresh [52] |
LOE II RCT, high estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 8 GC group: 8 |
Induced: 2 Intraperitoneal injections of l-arginine 2.5 g/kg, 1 h apart | Methylprednisolone 30 mg/kg/day PO 1 h post induction | None stated | CRP, histopathology 24 h after last L-arginine injection, score 0–3 (oedema, acinar cell degeneration, interstitial inflammation, haemorrhage) | NA | CRP results for GC group not shown | NA | NA | Authors claim less oedema, Inflammation, acinar cell degeneration & necrosis in GC group, results not shown |
| Duan et al. [63] |
LOE III Controlled trial, moderate estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 24 GC group: 21 |
Induced AP: sodium taurocholate 5% 1 mL/kg injected into the intracholangiopancreatic duct | Methylprednisolone 30 mg/kg IV 30 min after induction | 6 mL/kg/h SC saline | Mortality | NA | NA | NA | 4/24 model controls died, 0/21 in GC group died | NA |
| Wang et al. [36] |
LOE II RCT, moderate estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 25 GC group: 25 |
Induced AP: sodium taurocholate 5% 1 mL/kg injected into the intracholangiopancreatic duct | Dexamethasone 0.5 mg/kg IV 5 min after induction | None stated | Mortality, histopathology at 12 h, score 0–4 (oedema, acinar necrosis, haemorrhage, fat necrosis, inflammation, perivascular infiltrate) | NA | NA | NA |
Model controls: 42.9% GC group: 0% (P < 0.01) |
Haemorrhage & acinar necrosis less in GC group compared to model controls (P < 0.05) Results not shown |
| Liu et al. [53] |
LOE II RCT, low estimated risk of bias, small number of fairly well characterised rats in each group |
Rats |
Model controls: 16 GC group: 16 |
Induced AP: sodium taurocholate 5% 0.4 mL/kg injected into the intracholangiopancreatic duct | Glucocorticoid (type not stated) 20 mg/kg SC 1 h after induction | None stated | Mortality 24 h postinduction, histopathology at 3 h, 6 h and 12 h, score 0–3 (oedema, acinar necrosis, inflammatory infiltrate, haemorrhage, fat necrosis, perivascular inflammation) | NA | NA | NA |
Model controls 50% GC group 40%, Not significant |
Histopathology scores were lower in the GC group compared to model controls (P < 0.05) |
| Sha et al. [54] |
LOE II RCT, low estimated risk of bias, moderate number of well characterised rats in each group |
Rats |
Model controls: 24 GC group: 24 |
Induced AP: Taurocholate 40 g/L 1 mL/kg injected into the Intracholangiopancreatic duct | Dexamethasone 0.5 mg/kg IV after induction | None stated | Histopathology score 0–3 (oedema, acinar necrosis, haemorrhage, fat necrosis, inflammation, perivascular infiltrate and TEM at 3 h, 6 h and 12 h | NA | NA | NA | NA |
Histopathologic scores lower in GC group compared to model controls (P < 0.05) Ultrastructural changes were markedly less in the GC group |
| Kilic et al. [55] |
LOE II RCT, low estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 8 GC group: 8 |
Induced AP: cerulein injected IP hourly 5 times, given a total of 80 mg/kg |
Methylprednisolone 10 mg/kg IM twice hourly at 1 h after last cerulein injection | None stated | Histopathology at 20 h, score 0–3 (oedema), score 0–4 (inflammation, necrosis, vacuolisation) | NA | NA | NA | NA | Histopathology scores lower in GC group compared to model controls (P < 0.01) |
| Cui et al. [56] |
LOE II RCT, high estimated risk of bias, very small number of well characterised rats in each group |
Rats |
Model controls: 6 GC group: 6 |
Induced AP: sodium taurocholate 5% 1 mL/kg injected into intracholangiopancreatic duct |
Dexamethasone 1 mg/kg IP 3 h post induction | None stated | Histopathology at 24 h post treatment, no technique or results shown, only photos | NA | NA | NA | NA | Histopathological changes were alleviated in GC group according to authors |
|
Zhao et al. [64] Group 1 |
LOE III Controlled trial, high estimated risk of bias, small number of well characterised mice in each group |
Mice C57BL/6 |
Model controls: 34 Group 1: 20 Group 2: 15 |
Induced: l-arginine 2.5 mg/g injected IP twice with 1 h interval |
Dexamethasone 30 min post induction controls: No GC Group 1: GC 0.4 mg/kg IV Group 2: GC 4 mg/kg IV |
Geldanamycin 1 or 10 g/kg IV 10 min post induction | Mortality at 24 h, histopathology at 24 h, no technique or results shown, only photos | NA | NA | NA |
Model controls: 23.5% Group 1: 15% Group 2: 6.7% |
Pictures of histopathology are shown but not discussed |
|
Zhao et al. [64] Group 2 |
LOE III Controlled trial, moderate estimated risk of bias, small number of well characterised mice in each group |
Mice BALB/c |
Model controls: 34 Group 1: 20 Group 2: 15 |
Induced: l-arginine 2.5 mg/g injected IP twice with 1 h interval |
Dexamethasone 0.4 or 4 mg/kg IV 30 min post induction controls: No GC Group 1: GC 0.4 mg/kg IV Group 2: GC 4 mg/kg IV |
Geldanamycin 1 or 10 g/kg IV 10 min post induction | Mortality at 24 h, histopathology at 24 h, no technique or results shown, only photos | NA | NA | NA |
Model controls: 52.9% Group 1: 45% Group 2: 6.7% |
Pictures of histopathology are shown but not discussed |
Model controls: AP cases receiving no or standard (other) treatment, GC group: AP cases treated with GC in addition to treatment for model controls
*10 articles were found describing the same experimental study, the most relevant was used as reference here, the other 9 was full text excluded, see Fig. 1, a reference list is available at request
**2 articles were found describing the same experimental study, the most relevant was used as reference here, the other was full text excluded, see Fig. 1, a reference list is available at request
AP: acute pancreatitis; CRI: constant rate infusion; CRP: C-reactive protein; GC: glucocorticoid; h: hours; HC: hydrocortisone; IM: intramuscular; IV: intravenous; IP: intraperitoneal; NA: not applicable; NPO: nothing per os; Pred: Prednisolone; RCT: randomised controlled trial; SC: subcutaneous; TEM: transmission electron microscopy