TABLE 1

Characteristics of Included Studies

ReferenceGestational Age and Number of InfantsAge at Prognostic Testing and Prognostic VariablesAge at Follow-up and Abnormal OutcomeMain Findings
Jiang et al 201524GA: 34–36 wkaEEG 6–8 h: Background; cutoff between continuous and discontinuous normal voltage18 mo: Death, cerebral palsy,a MDI <70 or PDI <70Background: Sensitivity 0.81, specificity 0.92, PPV 0.64, NPV 0.96.
224 eligible → 201 included →170 at follow-upSleep–wake cycling; cutoff between marginally abnormal and noneSleep–wake cycling: Sensitivity 0.62, specificity 0.93, PPV 0.62, NPV 0.93.
Schwindt et al 201533GA: <30 wkaEEG ≤2 wk (separate data for ≤1 wk): Background; normal or abnormal24 mo: Death, mild impairment (MDI or PDI <85 and ≥70 or GMFCS grade ≥2), or severe impairment (MDI or PDI <70 and GMFCS grade ≥2)Combined aEEG score: 62.5% of the infants had a normal aEEG score, and out of these 76% had a normal neurologic outcome. Of the infants with an abnormal aEEG score, 24% had a normal neurologic outcome. P < .001.
156 included → 136 at follow-up → 50 had aEEG performed ≤1 wk → 48 included in analysisSleep–wake cycling; present or absent
Seizures; present or absent
Combined aEEG score (defined in article); normal or abnormal
Nunes et al 201431GA: 26–35 wkEEG days 2–14 (separate data for ≤1 wk): Background abnormalities, dysmaturity pattern, positive sharp waves, electrographic seizures; cutoff between present or absent for all variables12 mo: Developmental delay,b cerebral palsy,a or epilepsyBackground: RR 1.84 (95% CI 1.18–2.86) for developmental delay, P = .010.
47 included preterm infants (49 term controls) → 7 preterm infants had EEG performed ≤1 wk
Schumacher et al 201332GA: 24–31 wkEEG ≤3 d: Total absolute band power; cutoff at 1 SD below average median value2 y: MDI, PDI, or TMQ <85Day 1: δ-PP ranged from 0.60 to 0.67, δ-NP from 0.60 to 0.94, θ-PP from 0.29 to 0.57, θ-NP from 0.57 to 0.87, α-PP from 0.40 to 0.71, α-NP from 0.64 to 0.88, β-PP from 0.33 to 0.67, and β-NP from 0.60 to 0.84.
48 included → 41 at follow-up → 21 included in analysis
Hayashi-Kurahashi et al 20127GA: 22–33 wkEEG days 1–6: Acute and chronic stage abnormalities; cutoff between none and mild, and mild and moderate12–18 mo: Developmental quotientc <70 or cerebral palsyaAny grade of EEG abnormality: Sensitivity 0.61 (95% CI, 0.45–0.75), specificity 0.74 (95% CI, 0.68–0.80), PPV 0.27 (95% CI, 0.19–0.38), NPV 0.92 (95% CI, 0.88–0.95).
436 recruited → 333 included → 265 had EEG on days 1–6Moderate to severe grade of EEG abnormality: Sensitivity 0.41 (95% CI, 0.27–0.58), specificity 0.92 (95% CI, 0.88–0.95), PPV 0.46 (95% CI, 0.30–0.62), NPV 0.91 (95% CI, 0.87–0.94).
Le Bihannic et al 201230GA: <31 wkEEG ≤4 wk (separate data for ≤1 wk): Background and occurrence of abnormal features; normal or abnormal (disorganized, dysmature, or showing other abnormal features including seizures)1.5–6 y: Moderate or severe sequelaedSensitivity 0.83, specificity 0.88, PPV 0.91, NPV 0.79.
76 recruited → 61 included → 42 had EEG performed ≤1 wk
Wikström et al 201214GA: 22–30 wkaEEG <72 h: Background; cutoff between present or absent burst suppression, and present or absent continuity2 y: Death, survival with cerebral palsy,a MDI <70, PDI <70, blindness, or deafnessBurst suppression: Sensitivity 0.89, specificity 0.67, PPV 0.63, NPV 0.91.
54 included → 49 at follow-upSleep–wake cycling; cutoff between no and imminent cyclicityNo continuous activity: Sensitivity 0.89, specificity 0.53, PPV 0.54, NPV 0.89.
EEG <72 h: Seizures; present or absentNo cyclicity: Sensitivity 0.58, specificity 0.77, PPV 0.61, NPV 0.74.
Median interburst interval; cutoff at 6 sMedian interburst interval >6 s: Sensitivity 0.63, specificity 0.79, PPV 0.63, NPV 0.79.
Interburst percentage; cutoff at 55%Interburst percentage >55%: Sensitivity 0.59, specificity 0.90, PPV 0.77, NPV 0.79.
Klebermass et al 201115GA: 23–29 wkaEEG ≤1 wk: Background; normal or abnormal3 y: Death, cerebral palsy,a neurosensory impairment, MDI <85, or PDI <85Background: Sensitivity 0.81 (95% CI, 0.71–0.88), specificity 0.89 (95% CI, 0.76–0.96), PPV 0.93 (95% CI, 0.85–0.97), NPV 0.71 (95% CI, 0.58–0.83).
247 included → 143 at follow-up → 132 had aEEG performed ≤1 wkSleep–wake cycling; present or absentSleep–wake cycling: Sensitivity 0.72 (95% CI, 0.61–0.81), specificity 0.84 (95% CI, 0.71–0.93), PPV 0.89 (95% CI, 0.80–0.95), NPV 0.61 (95% CI, 0.48–0.73).
Seizures; present or absentCombined score: Sensitivity 0.87 (95% CI, 0.78–0.94), specificity 0.73 (95% CI, 0.58–0.85), PPV 0.84 (95% CI, 0.75–0.91), NPV 0.78 (95% CI, 0.66–0.89).
Combined aEEG score (defined in article); normal or abnormal
West et al 201128GA: 24–28 wkEEG <48 h (2-channel):18 mo: Death, MDI <70, or PDI <70Neurophysiologist assessment: Sensitivity 0.63 (95% CI, 0.41–0.84), specificity 0.93 (95% CI, 0.86–0.99), PPV, 0.75 (95% CI, 0.54–0.96), NPV 0.88 (95% CI, 0.80–0.96), LR+ 9.0 (95% CI, 3.2–24.6).
120 eligible → 84 recruited → 76 at follow-up Neurophysiologist assessment of interburst interval, continuity, seizures, amplitude, and developmental features; cutoff between mildly and moderately abnormalContinuity at 25-μV threshold: Sensitivity 0.58 (95% CI, 0.36–0.80), specificity 0.72 (95% CI, 0.60–0.84), PPV, 0.41 (95% CI, 0.22–0.60), NPV 0.84 (95% CI, 0.74–0.94), LR+ 2.0 (95% CI, 1.2–3.6).
Quantitative continuity assessment; cutoff at 25 and 50 μVContinuity at 50-μV threshold: Sensitivity 0.47 (95% CI, 0.25–0.70), specificity 0.77 (95% CI, 0.66–0.88), PPV, 0.41 (95% CI, 0.20–0.62), NPV 0.81 (95% CI, 0.71–0.92), LR+ 2.1 (95% CI, 1.1–4.1).
Kidokoro et al 201025GA: 27–32 wkaEEG <24 h: Background; no cutoff, all had discontinuous normal voltage12 mo: Developmental quotiente <70 or observed neurologic abnormalitiesAll 12 infants had discontinuous normal voltage on aEEG. Cyclicity was present in 8 infants who had favorable outcomes. One infant showed no cyclicity and had a favorable outcome, and 3 infants with absent cyclicity had abnormal outcomes.
108 recruited → 81 enrolled → 12 includedCyclicity; present or absent
Maruyama et al 200226GA: 27–32 wkEEG ≤1 wk: Acute stage abnormalities; cutoff between normality and 5 grades of abnormality18 mo: Mild, moderate, or severe cerebral palsyaSensitivity range days 0–2: 0.71–0.83 (decreasing after day 3).
318 recruited → 295 includedSpecificity range days 0–7: 0.80–0.96.
PPV day 2: 0.83 (lower at other time points).
NPV range days 0–7: 0.73–0.95.
Hellström-Westas et al 199129GA: 23–33 wkaEEG ≤1 wk: Continuity, burst suppression, sleep–wake cycling, inactivity, epileptiform activity; cutoff between present or absent for all variables2 y: Death or sequelaefContinuous activity: Sensitivity 0.67, specificity 0.94, PPV 0.91.
42 recruited → 31 includedSleep–wake cycling: Sensitivity 0.73, specificity 0.69, PPV 0.69.
Inactivity: Sensitivity 0.38, specificity 1.00, PPV 1.00.
Tharp et al 198127GA: <37 wkEEG ≤1 wk (first recording): Background and occurrence of abnormal features; normal or abnormal (moderate or major EEG abnormalities)2–7 y: Death or major or minor sequelaegNormal EEGs were present in all groups (1, normal outcome; 2, minor sequelae; 3, major sequelae; 4, dead) but were more common in groups 1 and 2. In group 1, 61% of the infants had normal EEGs. Moderately abnormal records were present in all groups and appeared to be of no value in predicting outcome. Major EEG abnormalities were present only in survivors with abnormal outcomes.
81 included → 33 had aEEG performed ≤1 wk
  • The results reported in the Main Findings column are in exact accordance with the reported results in the studies and are not based on additional information received from authors or calculations performed for the purpose of data synthesis. α, α activity; β, β activity; δ, δ activity; θ, θ activity; aEEG, amplitude-integrated EEG; ASA, acute stage abnormalities; GA, gestational age; GMFCS, gross motor function classification system evaluated by neurologic examination; MDI, mental developmental index evaluated by BSID-II; NPV, negative predictive value; PDI, psychomotor developmental index evaluated by BSID-II; PPV, positive predictive value; RR, relative risk; TMQ, total motor quotient evaluated by Peabody Developmental Motor Scales–2.

  • a Evaluated by neurologic examination.

  • b Evaluated by Denver-II test.

  • c Evaluated by Tsumori–Inage Infant Developmental Scale or Kyoto Scale of Psychological Development.

  • d Evaluated by a standardized interview and a complete clinical examination. If a deficit was detected, specific tests for oral language and cognitive skills development were carried out.

  • e Evaluated by Tsumori–Inage Developmental Questionnaire.

  • f Evaluated by neurodevelopmental examination.

  • g Evaluated by a general physical and neurologic examination, and for many also by a standard intelligence test (Brunet–Lézine or others).