SNAPPE-II: Score for Neonatal Acute Physiology with Perinatal Extension II

Clinical Overview

SNAPPE-II is a prospective, physiologically grounded severity-of-illness score designed to quantify acute illness burden in neonates during the first 12 hours of life. The score predicts in-hospital mortality and major morbidity in preterm and term infants admitted to neonatal intensive care, making it invaluable for standardizing severity assessment, comparing cohort outcomes, and facilitating family counseling.

What It Measures

SNAPPE-II integrates nine variables. The six physiologic items (SNAP-II component) are collected during the first 12 hours of NICU admission, while the three perinatal extension items (birth weight, SGA status, Apgar score) are known at birth:

1. Lowest systolic blood pressure (mmHg): Reflects perfusion adequacy and shock severity
2. Lowest body temperature (°C): Represents thermoregulatory stress and metabolic derangement
3. Highest P_aO2/F_iO2 ratio (assuming room air = 21%): Measures oxygenation efficiency; lower ratios indicate severe respiratory disease
4. Lowest serum pH: Reflects metabolic and respiratory acidosis severity
5. Presence of seizures (yes/no): Binary indicator of acute CNS dysfunction
6. Urine output (mL/kg/hour): Marker of renal perfusion and function
7. Birth weight (g): Intrinsic risk factor for mortality
8. Small for gestational age (SGA) status (yes/no): Associated with growth restriction complications
9. 5-minute Apgar score: Reflects immediate response to resuscitation

Each variable is assigned a point value on a scale weighted by its independent association with mortality. Higher scores indicate greater illness severity.

Why It Was Developed

The original SNAP score (1996) captured acute physiology but lacked incorporation of perinatal risk factors (birth weight, SGA, Apgar). Richardson and colleagues developed SNAPPE-II in 2001 to address this limitation, creating a unified score that incorporates:

• Physiological derangement (reflecting acute NICU illness)
• Intrinsic risk factors (reflecting inherent vulnerability)
• Perinatal events (reflecting delivery room response)

This comprehensive approach allows clinicians to distinguish between sicker-at-birth infants (high birth weight SGA with low Apgar) and initially well-appearing infants who deteriorate after admission. The score was prospectively derived and validated on >3,000 neonates across multiple centers.

Clinical Problem It Solves

Neonatal mortality prediction serves multiple critical functions:

1. Standardized severity comparison: Allows NICU quality improvement programs to compare case-mix adjusted outcomes (e.g., expected vs. observed mortality)
2. Family counseling: Provides data-driven mortality estimates to inform decision-making discussions in periviable and critically ill scenarios
3. Triage and resource allocation: Identifies highest-risk infants requiring intensive monitoring or advanced interventions
4. Research comparability: Enables prospective study cohorts to be matched by severity, reducing confounding by indication
5. Prognostication: Predicts not only mortality but also association with major morbidities (IVH, NEC, chronic lung disease)

Without standardized scoring, clinicians may rely on subjective impressions, leading to inconsistent counseling and suboptimal resource allocation.

When and Where to Use It

Optimal timing and populations:

• Physiologic variables captured within the first 12 hours of NICU admission (perinatal extension items known at birth)
• Applicable to all gestational ages (23 weeks to term) and birth weights
• Particularly useful in periviable counseling (22–25 weeks) and critical illness scenarios
• Less useful for stable term infants with mild respiratory disease (scores <10 often seen, predictive value limited)

Clinical settings:

• Delivery room/NICU admission assessment
• Multidisciplinary family discussions regarding goals of care
• Quality improvement committees comparing case-mix adjusted outcomes
• Research cohort stratification and matching
• Serial assessment to track illness progression (though designed for single time point)

Key Components

Each SNAPPE-II component adds dimension to the severity picture:

Interpretation Guide

How to Read Results: Risk Stratification Tiers

SNAPPE-II generates a point score (0–120), which maps to mortality risk according to the original derivation cohort (n=3,049 neonates):

Clinical Decision Points

Score ≤9 (Low Risk):

• Interpretation: Infant appears physiologically stable despite NICU admission
• Clinical actions:
• Standard NICU care appropriate
• Early advancement of feeds and developmental care
• Plan for discharge within standard timeline
• Family counseling: Reassure regarding short-term survival prognosis
• Routine follow-up sufficient

Score 10–19 (Mild-Moderate Risk):

• Interpretation: Some physiological derangement present; monitor closely for deterioration
• Clinical actions:
• Close vital sign monitoring; reassess SNAPPE-II if condition changes
• Interventions based on specific abnormalities (e.g., inotropes for hypotension, antibiotics if sepsis suspected)
• Anticipate potential complications but avoid excessive intervention
• Family counseling: "We expect your baby to improve with NICU care; we're watching closely for changes"
• Plan follow-up studies (cranial ultrasound at day 3–5, echocardiogram if needed)

Score 20–39 (Moderate-Severe Risk):

• Interpretation: Significant acute physiological compromise; outcomes uncertain
• Clinical actions:
• Implement aggressive support (mechanical ventilation, vasopressors, antibiotics, careful fluid management)
• Multidisciplinary rounds with realistic prognostication
• Escalate care as indicated; consider extracorporeal membrane oxygenation (ECMO) consultation if applicable
• Family counseling: This is the critical decision point. Families should understand:
• Survival is likely but not guaranteed
• Major morbidity risk (IVH, chronic lung disease, NEC) is elevated
• Consider offering "time-limited trial" of maximal support with reassessment at 48–72 hours
• Discuss values and goals; ensure consistent messaging across team
• Palliative care consultation if uncertainty about goals exists
• Frequent clinical assessment and score reassessment

Score 40–59 (Severe Risk):

• Interpretation: Critical illness with high mortality risk; many infants will not survive
• Clinical actions:
• Maximal organ support (intubation, vasopressors/inotropes, renal replacement if needed, surfactant)
• Intensive monitoring with serial assessments (daily or more frequently)
• Proactive family meetings with realistic prognosis:
• "We are very concerned about your baby's survival. We estimate less than a 50% chance of survival based on how sick they are right now"
• Discuss both best-case (survival with potential disabilities) and worst-case (death) scenarios
• Explore family values, religious beliefs, and goals
• Discuss what "comfort-focused care" would look like as an alternative
• Offer time-limited trial of 48–72 hours with clear criteria for reassessment and potential transition to comfort care
• Palliative care and spiritual care involvement strongly encouraged
• Document goals of care conversations in medical record

Score ≥60 (Critical/Likely Lethal Illness):

• Interpretation: Prognosis for survival is poor; mortality >50% even with maximal support
• Clinical actions:
• Reassess diagnosis and reversibility (consider sepsis workup, imaging, etc., but avoid futile testing)
• Proactive family meeting (within 24 hours) with realistic prognostication:
• "We do not believe your baby will survive despite our best intensive care efforts. We recommend focusing on comfort and family time"
• Offer options: Maximal support with clear reassessment window, comfort-focused care, or intermediate approach
• Avoid escalation unless family explicitly requests continued full support with understanding of realistic prognosis
• Facilitate holding, skin-to-skin care, baptism/blessing, memory creation (photos, handprints)
• Palliative care should be primary service in most cases

Common Pitfalls in Interpretation

Pitfall 1: Over-interpretation of a single score SNAPPE-II predicts population-level mortality but is less precise for individual infants. An infant with score of 25 has ~6% mortality risk, but this represents a range from ~4–11%. Some infants with SNAPPE-II 25 will die; others will survive without serious morbidity. Clinical course, response to initial therapy, and repeat scoring provide additional prognostic information.

Pitfall 2: Neglecting score reassessment SNAPPE-II is designed as a single time-point assessment (first 24 hours), but serial assessment can track illness trajectory. An infant scoring 45 at 6 hours who scores 15 at 36 hours (indicating response to therapy) has markedly improved prognosis compared to one scoring 50 at 36 hours (indicating deterioration). Reassessment provides dynamic risk stratification and guides escalation or de-escalation decisions.

Pitfall 3: Using SNAPPE-II as sole basis for goals-of-care discussions While SNAPPE-II provides crucial mortality estimates, it does not capture major morbidity risk, functional outcome, family values, or social context. A score of 35 in a well-resourced family with access to long-term follow-up may support more aggressive intervention than the same score in a resource-limited setting. Multidisciplinary discussions incorporating clinical, ethical, and social dimensions are essential.

Pitfall 4: Misinterpreting birth weight component SNAPPE-II incorporates birth weight as a risk factor, but this can lead to assumptions that all extremely low birth weight (ELBW) infants are "sicker." A 600 g infant with a SNAPPE-II score of 8 is NOT automatically sicker than a 2000 g infant with score 25. The score integrates the risk, but individual physiological status is paramount.

Pitfall 5: Assuming SNAPPE-II predicts neurodevelopmental outcome SNAPPE-II was derived and validated for mortality prediction. While there is association between high SNAPPE-II scores and major morbidities (IVH, NEC, BPD), the score does not directly predict neurodevelopmental disability or school performance. Families should understand this distinction: "We estimate a 20% chance your baby will not survive. Among survivors, some will have learning delays or other disabilities, but predicting this is difficult."

Evidence & Validation

Original Derivation Study

Richardson DK, Corcoran JD, Escobar GJ, et al. "SNAP-II and SNAPPE-II: Simplified Newborn Physiology Score and Expanded Physiology Score for Assessing Neonatal Illness." Journal of Pediatrics. 2001;138(1):92–100. DOI: 10.1067/mpd.2001.109608

Study design: Prospective cohort study of 3,049 neonates admitted to 35 NICUs across the US between 1997–1998

Subjects:

• Gestational age ≥23 weeks (mix of preterm and term)
• Birth weight ≥400 g
• NICU admission on day of life 0 or 1
• Diverse case mix (35% SGA, 18% <1000 g, 11% <750 g)

Outcome: In-hospital mortality; follow-up analysis included major morbidities (IVH grade 3–4, NEC stage 2–3, chronic lung disease)

SNAPPE-II components and point allocation (determined by logistic regression with mortality as outcome):

Results:

• Mortality in derivation cohort: 9.7% overall (22% in <750 g, 2% in >1500 g)
• SNAPPE-II score distribution: median 12 (IQR 5–24)
• Logistic regression: SNAPPE-II had strong independent association with mortality (odds ratio per 10-point increase: 1.85; 95% CI 1.72–1.99)
• C-statistic (area under receiver operating characteristic curve): 0.89, indicating excellent discrimination

Key Validation Studies

Validation 1: Original paper internal validation

Richardson et al. (2001) performed 10-fold cross-validation on the derivation cohort, demonstrating:

• C-statistic remained 0.88–0.89 in validation folds
• Mortality prediction was well-calibrated across risk groups
• Score remained predictive in subgroups (ELBW, periviable, SGA)

Validation 2: External cohort validation

Subsequent studies (Horbar et al., Vermont Oxford Network; Zeitlin et al., EPIPAGE study) validated SNAPPE-II across diverse populations:

• C-statistic: 0.85–0.88 in external cohorts
• Prediction remained accurate in single-center studies (n=100–500) and multicenter networks (n>5,000)
• Performance consistent across US, European, and Asian NICUs

Validation 3: Association with major morbidities

Original Richardson paper and subsequent research demonstrated:

• IVH grade 3–4: Adjusted odds ratio (AOR) 1.24 per 10-point increase in SNAPPE-II (p<0.001)
• NEC stage 2–3: AOR 1.18 per 10-point increase (p<0.001)
• Chronic lung disease (oxygen requirement at 36 weeks): AOR 1.15 per 10-point increase (p<0.001)

This indicates SNAPPE-II captures both immediate mortality risk and downstream morbidity burden.

Sensitivity and Specificity at Key Cutoffs

For mortality prediction (using score ≥40 as cutoff for "high risk"):

• Sensitivity (detecting future deaths): ~50% (varies by cohort; some score <40 but still die due to complications)
• Specificity (identifying survivors): ~95%

The score is better at ruling in risk (high specificity) than ruling out risk (moderate sensitivity). An infant with SNAPPE-II ≥40 has high mortality risk; an infant with SNAPPE-II <20 is still not guaranteed to survive.

Sample Sizes and Populations Studied

Limitations

1. Mortality-focused derivation: SNAPPE-II was derived to predict mortality specifically. While associations with morbidity exist, they are secondary outcomes. Prognostication for neurodevelopmental disability cannot be made from this score alone.

1. Temporal specificity: Score must be calculated within 24 hours of admission to maintain accuracy. Late-admission neonates (e.g., those transferred at 48 hours) do not fit the derivation criteria.

1. Missing physiological data: In ~5–10% of neonates, complete data may not be available (e.g., pH not drawn, BP not recorded). The original Richardson paper did not address handling of missing values; imputation strategies vary by implementation.

1. Cohort drift over time: The derivation cohort is from 1997–1998. Improvements in perinatal care, surfactant administration, antenatal steroids, and NICU practices may have shifted outcome distributions. A modern cohort might show different absolute mortality risks for the same SNAPPE-II score, though relative risk ordering likely remains valid.

1. Limited specificity for cause of death: A high SNAPPE-II score predicts increased mortality risk overall but does not distinguish between deaths from respiratory failure, sepsis, NEC, IVH, or other causes. Prognosis discussions must incorporate clinical diagnosis.

1. Does not predict long-term morbidity burden: Survivors with high SNAPPE-II scores may have varying disability severity. The score correlates with major morbidity presence but not severity or functional impact.

Worked Example

Clinical Scenario

Patient: Female, 24+4 weeks gestation, birth weight 580 g, singleton, non-SGA (weight appropriate for gestational age), vaginal delivery with maternal magnesium sulfate and antenatal corticosteroids.

Delivery room course: Intubated immediately in delivery room for respiratory distress. Received surfactant. Apgar scores: 1-minute = 3, 5-minute = 4, 10-minute = 6.

First 24 hours in NICU (time frame for SNAPPE-II calculation):

• Lowest systolic BP: 45 mmHg (at 2 hours of life; improved with normal saline boluses to 52 mmHg by 12 hours, but minimum is used)
• Lowest temperature: 34.2°C (initially; improved with servo control to 36.8°C by 4 hours, but minimum is used)
• Highest P_aO2/F_iO2: Initial ABG on FiO2 0.8 showed PaO2 52 mmHg; P_aO2/F_iO2 = 52/0.8 = 65
• Lowest serum pH: Initial VBG (umbilical cord) pH 6.95; subsequent ABG at 1 hour pH 7.08; lowest = 6.95
• Seizures: None observed
• Urine output (first 12 hours): No urine output recorded; estimated <0.1 mL/kg/hour (renal hypoperfusion in context of shock)
• Birth weight: 580 g
• SGA status: No (weight appropriate for 24 weeks)
• 5-minute Apgar: 4

Step-by-Step Calculation

Variable 1: Systolic BP (lowest = 45 mmHg)

• Reference table: <50 mmHg = 18 points
• Points: 18

Variable 2: Temperature (lowest = 34.2°C)

• Reference table: 35–35.9°C = 11 points (34.2°C falls in <35°C category)
• <35°C = 13 points
• Points: 13

Variable 3: P_aO2/F_iO2 (highest ratio = 65)

• Reference table: <100 = 15 points
• Points: 15

Variable 4: Lowest serum pH = 6.95

• Reference table: <7.0 = 19 points
• Points: 19

Variable 5: Seizures

• No seizures observed
• Points: 0

Variable 6: Urine output (lowest rate = <0.1 mL/kg/hour)

• Reference table: <0.5 mL/kg/hour = 10 points
• Points: 10

Variable 7: Birth weight = 580 g

• Reference table: 500–749 g = 13 points
• Points: 13

Variable 8: SGA status

• No (appropriate for gestation)
• Points: 0

Variable 9: 5-minute Apgar = 4

• Reference table: 4–6 = 10 points
• Points: 10

Total Score Calculation

Sum of all points: 18 + 13 + 15 + 19 + 0 + 10 + 13 + 0 + 10 = 98

Clinical Interpretation

SNAPPE-II score of 98 places this infant in the critical illness category with mortality risk >50% (likely 40–60% in modern cohorts, higher in the original 1997 data).

What this score tells us:

• This infant is one of the sickest neonates typically seen in NICU practice
• Expected survival is less than 50% even with maximal support
• Major morbidities (IVH, NEC, chronic lung disease) are very likely in survivors
• Neurodevelopmental disability risk is substantial

Clinical actions recommended at this SNAPPE-II level:

1. Immediate: Multidisciplinary family meeting within 6–12 hours (after parents have had time to process NICU admission)

• Realistic prognostication: "Your baby is extremely critically ill. We estimate less than 50% chance of survival despite all our efforts. Among those who survive, serious brain hemorrhage and long-term lung disease are common."
• Discuss what "full support" entails: intubation, medications to support blood pressure, possible blood transfusions
• Discuss alternative: "comfort-focused care" emphasizing pain relief, family time, memories

1. Clinical support: Aggressive organ support

• Continue mechanical ventilation
• Initiate vasopressor support (dopamine or dobutamine infusion) for hypotension
• Avoid hyperthermia; target temperature 36.5–37.5°C
• Serial blood gases to optimize pH and oxygenation
• Renal function monitoring; consider fluid restriction given oliguria
• Antibiotics covering maternal flora (group B streptococcus, gram-negatives)
• Surfactant if not already given; consider second dose if significant deterioration

1. Reassessment: Calculate SNAPPE-II again at 48 hours

• If score improves (e.g., BP normalizes, pH corrects, FiO2 requirements decrease), this signals physiologic recovery and supports continued aggressive care
• If score worsens or remains >80, this suggests poor prognosis and should prompt discussion of potential transition to comfort-focused care

1. Family support: Palliative care consultation; facilitate skin-to-skin holding if medically feasible; memory creation; spiritual care

1. Prognosis statement for discharge planning (if infant survives):

• Expected discharge after >100 days of hospitalization (at minimum)
• Likely home oxygen requirement
• High risk for neurodevelopmental follow-up abnormalities; baseline developmental assessment and early intervention referral essential
• Plan for ophthalmology (retinopathy of prematurity screening at 4 weeks)
• Optimize nutritional support for brain development

References

1. Richardson DK, Corcoran JD, Escobar GJ, Lee SK. SNAP-II and SNAPPE-II: simplified newborn illness severity and mortality risk scores. J Pediatr. 2001;138(1):92-100. doi:10.1067/mpd.2001.109608
2. Richardson DK, Gray JE, McCormick MC, Workman K, Goldmann DA. Score for Neonatal Acute Physiology: a physiologic severity index for neonatal intensive care. Pediatrics. 1993;91(3):617-623.