Abstract
Introduction
Neonatal jaundice is common, and clinically important hyperbilirubinemia requires timely recognition because excessive unconjugated bilirubin can cause acute bilirubin encephalopathy and kernicterus. For newborn infants at least 35 weeks of gestation, current American Academy of Pediatrics guidance emphasizes systematic bilirubin surveillance using transcutaneous bilirubin (TcB) or total serum bilirubin (TSB), interpreted according to postnatal age and neurotoxicity risk factors.[1]
The Kramer method, first described in 1969, grades visible jaundice according to its cephalocaudal progression from the face and neck toward the palms and soles.[2] Because it is rapid, noninvasive, and requires no equipment, Kramer staging remains attractive where laboratory testing or TcB devices are not immediately available. However, visual assessment is influenced by ambient light, examiner experience, skin pigmentation, and the broad overlap of bilirubin concentrations within visually defined zones. A systematic review of noninvasive bilirubin methods found direct visual assessment to be the least accurate and least reliable approach compared with laboratory bilirubin measurement.[3]
Recent studies continue to show variable performance of visual Kramer assessment across settings and populations.[4-7] In resource-constrained clinical environments, its most defensible role may therefore be as an initial bedside signal that prompts quantitative bilirubin measurement rather than as a stand-alone diagnostic or treatment tool. Data from Venezuela remain limited. A prior Venezuelan study reported a strong correlation between Kramer index and serum bilirubin,[8] but local performance may differ according to population, examiner technique, and the spectrum of jaundice severity.
The present study aimed to evaluate the association between Kramer visual jaundice grade and TSB among neonates with clinical jaundice at a public maternal-child hospital in Valencia, Venezuela. We hypothesized that higher Kramer grades would be associated with higher TSB concentrations.
Methods
Study design and setting
This was a single-center, cross-sectional clinical study with prospective data collection at Hospital Materno Infantil Dr. José María Vargas in Valencia, Carabobo, Venezuela. Enrollment occurred from January through June 2020. The report was prepared in accordance with STROBE recommendations for cross-sectional observational studies.
Participants and sampling
All newborns at the study institution during the study period constituted the target population. A nonprobability, purposive sample of 80 neonates with a clinical impression of neonatal jaundice was included. Neonatal jaundice was the only explicit inclusion criterion documented in the study protocol, and no additional exclusion criteria were specified. Recorded gestational ages ranged from 37 to 41 weeks. No a priori sample-size calculation was documented.
Data collection and variables
Data were collected using structured direct observation and review of maternal and neonatal clinical records. Investigators used a standardized study record form capturing maternal characteristics, neonatal gestational age, sex, birth weight, length, Apgar scores, potential blood-group incompatibility, Kramer grade, and serum bilirubin values. Clinical assessments were performed under supervision of the clinical tutor or an attending clinician. Study documentation did not record the number of visual examiners, inter-rater reliability, ambient-light standardization, the exact interval between Kramer assessment and blood sampling, or the laboratory assay platform used for bilirubin measurement.
Kramer staging was recorded on the standard five-zone cephalocaudal scale: grade 1, face and neck; grade 2, upper extremities and upper trunk; grade 3, lower abdomen and thighs; grade 4, legs; and grade 5, palms and soles. No grade 5 neonates were observed in this cohort. TSB was the prespecified bilirubin outcome for this manuscript.
Statistical analysis
Statistical analyses were performed using SPSS version 17.0. Categorical variables were summarized as counts and percentages. Quantitative variables were summarized using the mean and standard error (SE), with medians and ranges reported when available. TSB values were compared across Kramer grades using one-way analysis of variance (ANOVA), consistent with the prespecified analysis. Spearman rank correlation was used to evaluate the association between ordinal Kramer grade and continuous TSB. A two-sided P value <0.05 was considered statistically significant. Because no post hoc pairwise comparison was reported, the ANOVA result was interpreted only as an overall between-grade difference and not as evidence that every pair of Kramer grades differed significantly.
Ethics
The study protocol was reviewed through the research and bioethics process of Hospital Materno Infantil Dr. José María Vargas, Valencia, Venezuela. Separate institutional authorization by the Medical Directorate is documented in a signed approval dated February 13, 2020. The requirement for informed consent was formally waived by the approving body.
Results
Study population
Eighty neonates with clinically recognized jaundice were included, and Kramer grade and TSB were recorded for all 80 participants. Forty-five (56.3%) were male. Gestational age was most commonly 39 weeks (41.3%) or 38 weeks (36.3%). Mean birth weight was 3177.38 ± 36.18 g (mean ± SE), with a median of 3150 g and a reported range of 2300 to 4100 g. Baseline neonatal characteristics are summarized in Table 1.
Table 1. Baseline neonatal characteristics
Characteristic Value Gestational age, weeks 37 2 (2.5%) 38 29 (36.3%) 39 33 (41.3%) 40 14 (17.5%) 41 2 (2.5%) Male sex 45 (56.3%) Female sex 35 (43.8%) Birth weight, g, mean ± SE 3177.38 ± 36.18 Birth length, cm, mean ± SE 49.34 ± 0.18 Apgar score 9 at birth 72 (90.0%) Apgar score 10 at 5 min 73 (91.3%)
Values are n (%) unless otherwise indicated. SE, standard error.
Kramer grade and total serum bilirubin
Kramer grade 2 was the most frequent classification (32/80, 40.0%), followed by grade 1 (23/80, 28.8%), grade 3 (16/80, 20.0%), and grade 4 (9/80, 11.3%). No infant was classified as Kramer grade 5. Overall mean TSB was 10.69 ± 0.37 mg/dL. Mean TSB increased across Kramer grades, from 9.43 mg/dL at grade 1 to 14.06 mg/dL at grade 4. The overall ANOVA comparing TSB across grades was statistically significant (F = 5.44; P = 0.0019).
Table 2. Kramer grade distribution and total serum bilirubin
Kramer grade n (%) TSB, mg/dL, mean ± SE 95% CI for mean TSB Overall 80 (100%) 10.69 ± 0.37 9.95–11.43 Grade 1 23 (28.8%) 9.43 ± 0.69 7.99–10.86 Grade 2 32 (40.0%) 10.28 ± 0.48 9.30–11.25 Grade 3 16 (20.0%) 11.44 ± 0.86 9.60–13.27 Grade 4 9 (11.3%) 14.06 ± 0.98 11.81–16.31
Overall one-way ANOVA: F = 5.44, P = 0.0019. CI, confidence interval; SE, standard error; TSB, total serum bilirubin. No pairwise post hoc comparisons were reported.
Correlation analysis
Kramer grade demonstrated a moderate positive association with TSB (Spearman ρ = 0.42; P = 0.0006), indicating that higher visual grades tended to occur with higher TSB values. The magnitude of association was insufficient to support quantitative bilirubin estimation from Kramer grade alone.
Table 3. Correlation between Kramer grade and total serum bilirubin
Variables Spearman ρ P value Interpretation Kramer grade vs TSB 0.42 0.0006 Moderate positive association
Discussion
This study found a statistically significant but only moderate relationship between Kramer visual jaundice grade and TSB in 80 neonates with clinically recognized jaundice. Mean TSB increased progressively from Kramer grade 1 through grade 4, and the overall ANOVA was significant. Nevertheless, the 95% confidence intervals overlapped across adjacent grades, and the Spearman correlation coefficient was 0.42. These findings support a directional association but do not support use of Kramer grade as a precise surrogate for measured bilirubin.
The magnitude of correlation was lower than that reported in the prior Venezuelan study by Acosta-Torres and colleagues, which described a correlation of approximately 0.93.[8] Differences may reflect study population, disease spectrum, visual-assessment technique, examiner variation, or timing of serum measurement. Restricted spectrum may also have attenuated the present correlation because no infant reached Kramer grade 5 and only nine infants were classified as grade 4.
The present findings are consistent with the broader literature showing that visual jaundice assessment has variable diagnostic performance. Aprillia and colleagues reported clinically useful but imperfect sensitivity and specificity for Kramer examination,[4] while a Tanzanian study among neonates of Black African descent found only moderate reliability and cautioned against using Kramer assessment alone.[5] Other validation studies have likewise shown substantial variation in sensitivity and specificity depending on the threshold and population.[6] A 2021 review of noninvasive bilirubin methods concluded that direct visual assessment was the least accurate and least reliable modality when compared with laboratory bilirubin measurements.[3]
Current AAP guidance for infants at least 35 weeks of gestation relies on objective bilirubin measurement and interpretation by postnatal age and neurotoxicity risk factors, rather than visual staging alone.[1] More recent health-system evidence also suggests that adding TcB to Kramer-based assessment can substantially reduce the need for blood sampling while maintaining clinical surveillance.[7] Thus, in environments where TcB is available, an objective noninvasive measurement is preferable to visual estimation. Where TcB is unavailable, Kramer staging may still be useful as a rapid bedside cue to prioritize infants for TSB measurement, but it should not independently determine treatment.
This study is relevant to resource-limited settings because the original clinical environment had constrained access to bilirubin testing. A low-cost visual tool may help clinicians recognize the need for escalation, but the moderate association observed here underscores the risk of relying on skin appearance to infer a numerical bilirubin concentration. The appropriate clinical message is therefore not to replace laboratory testing with Kramer staging, but to use visual assessment as one component of a structured pathway that leads to timely quantitative measurement.
Limitations
The study has several limitations. First, it was a single-center nonprobability purposive sample of only 80 jaundiced neonates, which limits generalizability and introduces selection and spectrum bias. Second, no a priori sample-size calculation was documented, and no Kramer grade 5 infants were included, restricting assessment of the highest visual zone. Third, study documentation did not record the exact postnatal age in hours at visual assessment or TSB sampling, the interval between the two measurements, the bilirubin assay platform, examiner blinding, lighting standardization, or inter-rater reliability. These factors are important because bilirubin concentration changes over time and visual assessment is observer dependent. Fourth, no prespecified TSB threshold, sensitivity, specificity, likelihood ratios, or receiver-operating-characteristic analysis was performed; consequently, this should be interpreted as an association study rather than a diagnostic-accuracy validation study. Fifth, no multivariable analysis was performed to account for potential confounding by postnatal age, hemolytic risk, feeding-related factors, or other neonatal characteristics. Finally, data collection occurred during early 2020 and was disrupted by pandemic-related institutional-access limitations.
Conclusions
Kramer visual jaundice grade was moderately associated with TSB in this Venezuelan cohort, and mean TSB increased across grades 1 through 4. However, the overlap in bilirubin ranges and the moderate correlation indicate that Kramer staging should not be used to infer an exact bilirubin concentration or to make treatment decisions without objective measurement. In resource-constrained settings, Kramer assessment may serve as a preliminary bedside trigger for TcB or TSB testing rather than a substitute for quantitative bilirubin assessment.
References
- Kemper AR, Newman TB, Slaughter JL, et al. Clinical Practice Guideline Revision: Management of Hyperbilirubinemia in the Newborn Infant 35 or More Weeks of Gestation. Pediatrics. 2022;150(3):e2022058859. doi:10.1542/peds.2022-058859.
- Kramer LI. Advancement of dermal icterus in the jaundiced newborn. Am J Dis Child. 1969;118(3):454-458. doi:10.1001/archpedi.1969.02100040456007.
- Okwundu CI, Saini SS. Noninvasive methods for bilirubin measurements in newborns: a report. Semin Perinatol. 2021;45(1):151355. doi:10.1016/j.semperi.2020.151355.
- Aprillia ZR, Gayatri D, Waluyanti FT. Sensitivity, specificity, and accuracy of Kramer examination of neonatal jaundice: comparison with total bilirubin serum. Compr Child Adolesc Nurs. 2017;40(sup1):88-94. doi:10.1080/24694193.2017.1386975.
- Dionis I, Chillo O, Bwire GM, et al. Reliability of visual assessment of neonatal jaundice among neonates of black descent: a cross-sectional study from Tanzania. BMC Pediatr. 2021;21:383. doi:10.1186/s12887-021-02859-x.
- Mumtaz S, Shafiq MF, Khan KA, Hussain M, Rahman A, Shafiq S. Diagnostic accuracy of Kramer visual assessment of jaundice in term neonates taking total serum bilirubin levels as the gold standard. Pak Armed Forces Med J. 2022;72(6):1953-1956. doi:10.51253/pafmj.v72i6.6252.
- Lim XJ, Ambigapathy S, Leong EL, et al. Comparing Kramer’s rule with transcutaneous bilirubin vs Kramer’s rule only in reducing total serum bilirubin sampling among neonates with jaundice. BMC Pediatr. 2025;25(1):169. doi:10.1186/s12887-025-05423-z.
- Acosta-Torres S, Torres-Espina M, Colina-Araujo J, Colina-Chourio J. Utilidad diagnóstica del método de Kramer para la detección clínica de la hiperbilirrubinemia neonatal. Invest Clin. 2012;53(2):148-156.