Evidentiary table 1
PICO Questions: Is there an increased risk of adverse mechanical or infectious events in adult patients undergoing EVD insertion outside the OR? In adult patients undergoing EVD insertion, does the risk of adverse events vary depending on the training, procedural experience or specialty of the clinician performing the procedure?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
13 Gardner 2009 / 188 / 2, 3, 4, 6 / CS-R / Post-procedural hemorrhage following EVD insertion in OR vs outside OR / No significant difference in post-procedural hemorrhages whether the EVD was placed in the
OR versus ICU / low
21 Foreman 2015 / 150 / 1-6 / CS-R / To compare safety and accuracy of placement between EVDs placed in the ICU versus OR / Complications of hemorrhage, infection, and non-functional drains may be mitigated by ventriculostomy placement in the OR. / low / Patients who underwent ventriculostomy placement in the ICU differed in important ways (i.e. indication for placement and the administration of pre-procedure prophylactic antibiotics) from patients treated in the OR
12Kakarla 2008 / 346 / 1, 2, 8 / CS-R / To study the safety and accuracy of ventriculostomy by neurosurgical trainees / Neither the resident's training experience nor the side of placement seemed to affect accuracy / moderate
8 O’Neill 2008 / 29 / 1, 2, 5, 6, 7 / CS-R / To evaluate accuracy and complication rates of EVD insertions by non-neurosurgeons / No significant differences in complication rates relative to neurosurgical series when EVD performed by non-neurosurgeons / low / Mixed population of EVD and fiberoptic ICP monitors
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Evidentiary table #2
PICO Question:In adult patients undergoing EVD insertion, does the risk of adverse events vary depending on the training, procedural experience, or specialty of the clinician performing the procedure?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
33 Pollock JR 2001 / 728 / Craniotomy, shunt, Spinal surgery, stereotactic bx / CS-R / Surgical AEs in relation to # assisted cases / Inverse relation between number of assisted cases and rate of AEs / low
35Ehtisham 2009 / 29 / 1, 2, 3, 5, 6, / CS-R / EVD-related AEs / 1 infection (3.4%), 5 catheter tract hematomas <5cc (20.7%) when EVD inserted by NI / low
12 Kakarla 2008 / 346 / 1, 2, 3, 4, 8 / CS-R / Accuracy of placement, AEs / Supervised junior trainee AE rates same as experienced residents / low
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
AEs=Adverse Events, NI=neurointensivist
Evidentiary table #3
PICO Question: What is the risk of bleeding with insertion of an EVD?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
#12 Kakarla 2008 / 346 / 1, 2, 3, 8 / CS-R / Placement accuracy and AEs / 4 symptomatic bleeds / low / No risk factors for bleeding were identified
#35Ehtisham 2009 / 29 / 1, 2, 3, 4, 5, 6, 8 / CS-R / Placement accuracy and AEs / 0 symptomatic bleeds / low / All procedures by neurointensivist.
#11 Maniker 2006 / 160 / 1, 2, 3, 4, 5, 7 / CS-R / Hemorrhagic
AEs / 4 symptomatic bleeds / low / Non-significant increase in bleeding with smaller diameter catheter
#13 Gardner 2009 / 188 / 1, 2, 3, 4, 6 / CS-R / Hemorrhagic AEs / 77 (41%) bleeds / low / MR showed many bleeds, most small. 1 required surgery. Overall symptomatic bleeds unknown
#44Naff 2011 / 48 / 1, 3 / R / AEs / 28% / low (for symptomatic bleeds) / rtPA dose escalation.
# 43Dey 2015 / 250 / 1, 3 / R / Bleeding, infection / 16.8% bleeds, 2.4% symptomatic / low / CLEAR-III ongoing.
Blinded for rtPAvs placebo reduced evidence quality for bleeding outcome
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial, AE=adverse events
Evidentiary table #4
PICO Question:What procedural factors are associated with a decreased risk of catheter malposition?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
12 Kakarla 2008 / 346 / 1,2,3,4,8 / CS-R / Devised grading system for accuracy of positioning / 77% ideal
10% functional
13% malposition / low / Freehand placement
47Patil 2013 / 109 / 1,2,3,4,5,7 / CS-R / Malposition rates / 79% ideal;
TBI indication less often associated with ideal placement / low / Freehand placement
48Bergdal 2013 / 147 / 1,2,3,4,5 / CS-R / Correct EVD placement / Bolted EVDs more accurate, and associated with lower reoperation rates. / low / Bolt with freehand placement or freehand alone
49Huyette 2008 / 97 / 1,2,3,4,5 / CS-R / Catheter positioning / 22.4% malposition rate, only 56% in target location / low / Freehand at bedside
50 Abdoh 2012 / 66 / 2,3 / CS-R / Catheter positioning / 4% extraventricular, 20% contralateral / low / Freehand bedside
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
AEs=Adverse Events, NI=neurointensivist
Evidentiary table #5
PICO Question: In adult patients requiring EVD, what is the optimal method and timing of VTE prophylaxis?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
Risk of Hemorrhage with Prophylaxis
#88 Macdonald
2003 / 100 / Craniotomy / Pilot single center RCT / DVT/PE within 1 month
Postoperative intracranial hemorrhage / No significant differences in rates of DVT (0% vs 2%) or hemorrhage (2% vs 4%) in patients between heparin or LMWH VTE prophylaxis starting at the time of surgery / Low / Underpowered to detect small differences in clinically significant endpoint
#14
Tanweer
2013 / 99 / 1,2,3,4,5,6 / CS-R / EVD-associated hemorrhage / No significant differences
in the incidence of new hemorrhages (19.6% vs 16.3%), radiographically-significant hemorrhages (1.8% vs. 7.3%) and clinically significant hemorrhages (1.8% vs.3.6%) between patients starting heparin prophylaxis within 24 hours or after 24 hours / LOW
Grad[14]uated Compression Stockings (GCS)
#76
CLOTS trial 1
2009 / 2518 / Ischemic stroke, hemorrhagic stroke / Multicenter RCT / Proximal DVT / No significant difference in DVT with thigh length stockings.
Significant increase in skin breakdown with stockings (odds ratio 4.2, 95% CI 2.4-7.3) / Moderate / Establishes potential harm from graduated compression stockings in comparable population
#62
Agnelli 1998 / 307 / 5
(patients undergoing elective neurosurgery) / Multicenter RCT / Symptomatic PE or any DVT within 8 days / Decreased rate of DVT in patients receiving LMWH + GCS compared to GCS alone (32% vs. 17%; p<0.01)
No significant difference in major or minor bleeding between groups / High / GCS alone associated with high rate of DVT, which is substantially reduced by addition of LMWH prophylaxis
Inferior Vena Cava (IVC) Filters
#84
Decousus
1998 / 400 / Patients with proximal DVT / Multicentre RC: pts randomized to anticoagulation + IVC filter OR anticoagulation alone / PE within first 12 days of randomization / No difference in PE at 2 years with IVC added to anticoagulation
At 8 years follow up:
Decrease in PE with IVC filter (6.2% vs. 15.1%; p<0.01)
Increased DVT with IVC filter
(35.7 vs. 27.5%; p=0.04)
No difference in mortality / Low / True efficacy in reducing PE in neurosurgical patients not clear as all patients in this study received anticoagulation.
However utility in primary prophylaxis may be limited by associated increase in subsequent DVT
Intermittent Pneumatic Compression
#78
Dennis
2013 / 2876 / Ischemic stroke, hemorrhagic stroke / Multicenter RCT / Proximal DVT within 30 days / Decreased VTE with IPC (adjusted OR 0·65 [95% CI 0·51–0·84])
Increased skin breakdown with IPC (3% vs. 1 %; p<0.01)
Trend towards decreased mortality with IPC (11% vs. 13%; p=0.06) / Moderate / Establishes efficacy of IPC at preventing VTE in comparable population; with associated small risk but significant risk of skin breakdown.
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Evidentiary table 6
PICO Question: InfectionIntroduction
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
#97
Chan 1988 / 34 / 3,6 / CS-P / VRI / No infections / VL / Unique, experimental EVD system
#98
Omar 2010 / 87 / 1,2,3,4,5,8 / CS-P / VRI / 32% VRI total, 52% with VRI after 10 days / VL / VRI definition: (+) culture AND (+) gram stain AND CSF changes(WBC>11, glucose<25, protein>40)
#101
Mayhall 1984 / 172 / 2,3,4,5 / CS-P / VRI / Risk(life table) of infection: D#5=9%, D#8=21%, D#10=37%, D#11=42% / VL / VRI definition: (+) culture 24h after insertion
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Endpoint: VRI=Ventriculostomy related infection
Quality: VL=very low, L=low, M=moderate, H=high, VH=very high
Evidentiary table 7
PICO Question:In adult patients with an EVD, does the risk of infection increase with duration of placement?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
#31Arabi 2005 / 84 / 4,7,8 / CS-P / VRI / 1.2 OR total, OR (< 7 days) 2.82; OR >7 days 1.18 / VL / VRI Definition: Primary: (+)culture of known pathogen, secondary: "contaminant" with CSF or clinical meningitis OR CLINICAL only (CSF and pt changes)
#102
Korinek 2005 / 175 / 1,2,3,4,8 / CS-P, CS-R / VRI / EVD >5d vs EVD>10d not a risk factor for infection. Days mean+/- SD VRI 12.5+/- 11, no-VRI 9.4+/-7 / VL / VRI Definition: Fever >38.5C, (+) a positive CSF culture associated with CSF pleocytosis (>15 cells) CSF:plasma glucose ratio of less than 0.5
#103
Scheithauer 2009 / 1333 / NA / CS-P / VRI / No significant difference in VRI in EVDs in less than or more than 9 days / L / VRI Definition: CDC definition + CSF cellular and chemical changes + Clinical changes
#104
Arif 2012 / 104 / 1,2,3 / CS-R / VRI / EVD> 8 days 16/19 VRI; EVD<4 days 3/19 VRI / VL / VRI Definition: (+) CSF culture. In the absence of positive culture, infection was defined as greater than 50% PMN on CSF count with a minimum of 50 cells or CSF glucose less than 15mg/100ml
#105
Bota 2005 / 638 / 1,2,3,4,6,7 / CS-P / VRI / No infxn until d#3, linear increase d3-10, sharp cutoff of infections after D#10 / L / VRI Definition: Per Lozier, et al. [91]
#106
Camacho 2011 / 119 / 1,2,4,5 / CS-P / VRI / Catheter days no-VRI 7 (1-33); VRI: 10.3(4-26) / VL / VRI Definition: CDC, Horan, et al. [92]
#107
Chi 2010 / 197 / 1,2,3,4,5,8 / CS-R / VRI / Mean days-VRI:20+/-7.8d, no-VRI 14.7+/-10d / VL / VRI Definition: Clinical suspicion + culture of a pathogen from CSF (except d#1)
#108
Lyke 2001 / 157 / 1,4,5,7,8 / CS-P / VRI / VRI: 8.5d (2-23), no-VRI: 5.1d (1-23). 73% of VAI developed after 6d / VL / VRI Definition: Recognized pathogen OR skin flora and 1 or more criteria:(+) gram stain, low glucose, high protein, CSF PMN>10, excluded those with only (+) CSF parameters and (-) culture
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Endpoint: VRI=Ventriculostomy related infection
Quality: VL=very low, L=low, M=moderate, H=high, VH=very high
Evidentiary table 8
PICO Question:In adult patients, do prophylactic systemic antimicrobials reduce the incidence of VRI? Should a periprocedural or duration regimen be used?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
Arabi 2005 [31] / 84 (58 for comparison) / 4,7,8 / CS-P / VRI / Periprocedural:
Cefazolin OR Ceftriaxone OR Cefuroxime.
Antibiotic use associated with lower VRI rate; infection developed in 7 of 58 (12%) with antibiotics and in 12 of 41 (29%) when no antibiotics used (P = .03). / VL / VRI Definition: Primary: (+)culture of known pathogen, secondary: "contaminant" with CSF or clinical meningitis OR CLINICAL only (CSF and pt changes)
#106
Camacho 2011 / 119 / 1,2,4,5 / CS-P / VRI / Periprocedural:
1st or 2ndor 3rdor 4thgeneration cephalosporin. Use of antibiotics increased risk of VRI (non-significantly p=.22) with an OR of 1.8 / VL / VRI Definition: CDC, Horan, et al. [92]
#109
Dellit 2014 / 721 / 4 / CS-R / C. difficile colitis / Duration vsPeriprocedural:
Cefazolin
No significant difference in VRI (p=.29). Significant drop in C. difficile when changed to periprocedural (5.4% to 2.4%, p=.04) / VL / VRI definition: (+) CSF culture
#101
Mayhall 1984 / 172 / 2,3,4,5 / CS-P / VRI / Periprocedural:
Nafcillin 4 doses. No significant difference between antibiotic and no antibiotic group (p=.09) / VL / VRI definition: (+) culture 24h after insertion
#102
Alleyne 2000 [ / 308 / 4,5,8 / CS-P / VRI / Duration vsPeriprocedural.
Cefuroxime. No significant difference in VRI rate between duration (3.9%) and periprocedural (3.8%) / VL / VRI definition: (+) CSF culture
#111
Poon 1998 / 228 / 4,5,8 / R / VRI / Duration (D) vsPeriprocedural (P)
Amp/Sul +Aztreonam (D) vs Amp/Sul (P)
Significant reduction in VRI in D cohort (p=.01), but more resistant organisms and higher mortality in D cohort. / L / VRI definition:
Not stated
#112
Saini 2012 / 42 / 1,2,3,4,5,8 / R / VRI / Duration (D) vsPeriprocedural (P).
Ceftazadime.
No significant difference in VRI between D (6.67%) and P(7.4%) / VL / VRI definition:
(+) CSF or EVD tip cultures. CSF drawn on D# 0,1,3,5
#113
Wong 2006 / 255 / 4,5,8 / R / VRI / Duration: Two antibiotic regimen. Amp/Sul +aztreonamvsCefepime. Cefepimecohort with 6% VRI, Amp/Sul+aztreonam with 12% VRI / VL / VRI Definition:At least one of the following criteria must have been met: (1)(+) Culture and/or (2) fever (>38 °C) and any of the following: (a) Increased white cells (>50% PMN), increased protein, and/or decreased glucose (<15 g/dL) or (+) CSF gram stain. All CSF infections within 3 months after EVD insertion considered to be related.
#114
Blomstedt 1985 / 52 / NA / R / VRI / Duration: TMP-SMX vs. placebo. No difference in VRI (1 from each group) / VL / VRI Definition: NA
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Endpoint: VRI=Ventriculostomy related infection
Trimethoprim/sulfamethoxazole=(TMP/SMX)
Ampicillin/Sulbactam=Amp/Sul
Quality: VL=very low, L=low, M=moderate, H=high, VH=very high
Evidentiary table #9
PICO Question: In adult patients with an EVD, does the use of antimicrobial-impregnated catheters reduce the incidence of VRI?
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
#115
Zabramski 2003 / 306 / 2, 4,8 / R / VRI
MLR for prediction of VRI / ABX 1.3% vs. S 9.4% (p=0.002)
ER placement (OR 7.3; p=0.035), Presence of other infection (OR 5.1;p=0.047), ABX catheter (OR 0.14, p=0.014) / H / VRI defined as positive CSF culture (i.e. organism grew on two different media)
#116
Wong 2010 / 184 / 9, 4 / R / VRI / ABX 51% vs. S 57% (OR 1.3; 95% CI 0.7-2.2) / M / VRI defined as a positive CSF culture and CSF white blood cell count > 10/mm3CSF protein > 0.8 g/l, and CSF serum glucose ratio of < 0.4.
Open label trial.
Possible type 2 error
#117
Pople 2012 / 484 / 1, 10, 11 / R / Proven VRI
Suspected VRI
Duration of time to suspected infection / ABX 2.3 vs. S 2.8% (p=1.0)
ABX 17.6% vs. 20.4% (p=0.504)
ABX 8.8±6.1 days vs S 4.6±4.2 days (p=0.002) / M / Open label trial.
Proven VRI was defined as positive gram stain and positive culture grown in agar. Suspected infection was defined as culture positive but not seen on gram stain or culture negative but positive gram stain or CSF leukocytosis with a white blood cell/red blood cell count of > 0.02
#118
Harrop 2010 / Period 1 = 327 EVD
Period 2 = 281 EVD
Period 3 = 195 EVD / 8 / POC / VRI / Period 1 6.7% vs. Period 3 1.0% (p=0.005) / L / VRI defined as 2 positive CSF cultures and a concurrent increase in the CSF white blood cell count.
ABX catheters introduced in period 3; baseline = period 1
#119
Muttaiyah2010 / 120 patients and 986 CSF samples / 8 / POC compared to historical controls / VRI / ABX 5% vs. C 15% (P=0.0627) / VL / VRI defined as a positive CSF culture and CSF leukocytosis > 5 x 106 / L, CSF leukocytosis were corrected for erythrocytes in the CSF by subtracting 1 leukocyte for every 1000 erythrocytes in the CSF.
#120
McLaughlin 2011 / 75 / 2 / POC / VRI
Time to Infection / 9.3%
Median of 16 days / VL / No Comparison group.
VRI defined as organism isolated via culture and fever, headache, nuchal rigidity, meningeal signs, irritability AND antimicrobial treatment AND any of the following: positive gram stain, increased white blood cells, elevated proteins and/or increased decreased CSF glucose, organism isolated from blood culture.
#121
Mikhaylov 2014 / 145 / 10, 1, 2, 9, 4, 5, / Retro / VRI / ABX 4% vs. 10% (p=0.193) / VL / VRI defined as a positive CSF culture. Possible type 2 error
#122
Keong 2012 / 278 / 1, 2, 4, 12 / R / VRI / Silver 12.3% vs. 21.4% (p=0.0427) / H / VRI was defined as organisms seen on microscopy or isolated by culture or clinical meningitis requiring treatment.
#123
Lackner 2008 / 39 / 1,2,9, 5, / POC compared to historical controls / VRI / Silver 0% vs. S 25% (p<0.05) / VL / Possible Type 1 error.
VRI defined as positive culture.
#124
Fichtner 2010 / 164 / 1,2,4,8 / Retro / VRI / Silver 18.9% vs. S 33.7 (p=0.04) / L / VRI defined as positive culture or colonization of catheter tip with microorganism or raised liquor cell count > than 4 cells/mcl.
#125
Lajcak 2013 / 403 patients with 529 EVD / 1,2,4,5 / Retro / VRI / Silver 6.1% vs. S 13.8 (p=0.003) / VL / VRI defined as positive CSF culture.
#126
Winkler 2010 / 61 / 2 / R / VRI / Silver 10% vs. ABX 18% (p=0.71) / L / Confirmed VRI defined as leukocyte count per visual field > 200 and clinical signs of CSF infection present and microorganisms identified in the CSF or catheter tip. Probable VRI defined as leukocyte count per visual field > 100 and/or one or cmore of the confirmed VRI items.
Possible type 2 error
#127
Lemcke 2012 / 95 / 1,2,4,10 / Retro / VRI / Silver 9.4% vs. ABX 6.5% vs. S 15.6 (p>0.05) / VL / VRI defined as the identification of microorganisms in the CSF.
Possible type 2 error
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown 9=stroke 10=hydrocephalus 11=Edema 12=AVM 13=SDH
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial, POC = Prospective Observational Cohort, Retro = retrospective
Endpoint: VRI=Ventriculostomy related infection, MLG=multivariate logistic regression
Findings: ABX=antibiotic impregnated catheter, S=Silicone catheter, C=Control
Quality: VL=very low, L=low, M=moderate, H=high, VH=very high
Reference[ #] / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
#133
Pfausler 2003 / 10 / 1, 4, 6 ( Staphylococcal VRI) / R – Vent only Vs IV only / -CSF bacteriologic clearance.
-CSF vancomycin Levels / Higher CSF vancomycin level with intraventricular compared to intravenous with no difference in CSF bacterial clearance / VL / The study, though an RCT, was underpowered
#134
Remes 2013 / 34 / 1,2,5,6,7 / CS Vent+ IV Vs IV only / -Time to CSF bacteriologic clearance.
-Modified Rankin scale.
-Attributable death.
-Adverse effects. / Earlier time to CSF clearance and clinical outcome ( see text for details) / VL / Underpowered
#129
Wilke 2013 / 48 / 1, 2,3,5, 7 / CS-R
Vent+ IV Vs IV only / -Time to CSF bacteriologic clearance.
-Time to normalization of CSF WBC count.
-Length of stay.
-Length of IV. / All the outcomes were significantly better with vent+IV ( see text for details) / VL / Retrospective study with a scope for bias and confounding
Evidentiary table 10
PICO Question: Are additionalintra-ventricular antimicrobials effective for the treatment of VRI as compared to intravenous antimicrobials alone?
PICO Question: Are additionalintra-ventricular antimicrobials effective for the treatment of VRI as compared to intravenous antimicrobials alone?
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown
Study Design: CS-R=Case series cohort-retrospective, CS-P=Case series cohort-prospective, R=randomized trial
Endpoint: VRI=Ventriculostomy related infection
Quality: VL=very low, L=low, M=moderate, H=high, VH=very high
Interventions: Vent- Intraventricular antimicrobials, IV: intravenous antimicrobials
Evidentiary table 11
PICO Question: In adult patients requiring an EVD, does routine CSF sampling increase EVD-related infections as compared to maintaining a closed system with sampling of CSF only when clinically indicated?
Ref No. / Author/Year / Patient number / Underlying Clinical Condition(s) / Study Design / Endpoint / Findings / Quality / Comments
31 / Arabi/
2005 / 84 / 1, 4, 5, 6 / CS-R / Frequency of CSF sampling / Found no difference in VRI* – CSF daily samples vs. CSF samples when indicated / Very low / Single center, small number of patients, retrospective review
135 / Stenager/
1986 / 85 / 1, 2, 3, 4, 5 / CS-P / Frequency of manipulations was monitored; included CSF samples, irrigation when occluded, change of EVD / Found no difference in VRI associated with number of catheter manipulations / Very low / Single center, small number of patients, CSF sampling included with other manipulations
136 / Kitchen/
2011 / 133 / 2 / CS-R / Frequency of CSF sampling; compared control group with sampling when indicated vs. research group is frequent sampling / Frequency of VRI was significantly lower in research group with sampling group / Very low / Single center, retrospective review, control group very high VRI rate of 52.1%; comparison of the effect of sampling frequency on VRI in research group was complicated by implementation of an EVD management bundle
137 / Williams/
2011 / 382 / 2, 4 / CS-P / Historical control group – daily CSF sampling vs. every 3rd day CSF sampling / Switch to every 3rd day sampling was associated with a significant decrease in VRI (p = 0.02) / Very low / Confounding influence was use of AI-EVD* in 75% of patients in prospective study arm of every 3rd day CSF sampling
138 / Williamson/
2014 / 420 / Not provided / CS-R / Monitored frequency of CSF sampling / Univariate and multivariable analysis modeling found CSF sampling was associated with a significantly increased risk of VRI / Low / Single center, frequency of CSF sampling was monitored and not specified
Key: Underlying Clinical Condition 1=ICH 2=SAH 3=IVH 4=TBI 5=tumor 6=infection 7=nontraumatic 8=unknown