Total Macroscopic Maceration Score (TMMS): A Proposal for the Assessment of Maceration in Stillbirths

Introduction

Stillbirths remain a worldwide concern and are considered key indicators of the quality of care during pregnancy and childbirth. Since the World Health Organization (WHO) 2014 initiative “Every newborn: an action plan to end preventable deaths (ENAP),” ¹ reducing stillbirth rate has represented a global priority, requiring coordinated and targeted interventions at subnational levels. ² According to the WHO, “stillbirth” refers to the birth of baby with no signs of life at 22 or more weeks of gestation; however, for international comparisons, only deaths after 28 weeks of pregnancy are typically counted. ² The terms often overlaps with “fetal death,” which is “irrespective of the duration of pregnancy.” ³

Forensic pathology plays a central role in the multidisciplinary investigation of perinatal deaths, especially in the context of national mortality audit systems recommended by ENAP, ensuring essential measures to track the quality of the healthcare. ² A complete post-mortem examination of the fetus and placenta is essential to determine the cause and timing of death in stillbirths, identify contributing causes or avoidable factors, and provide information relevant to the definition of potential medical liability.^4-9

In order to understand liability profile, the estimation of the time of death in stillbirths is particularly crucial and could offer valuable insights into the potential for effective intervention and/or the fetus's likelihood of survival. ¹⁰

Nevertheless, estimating the timing of onset of fetal distress and the time of intrauterine death remains controversial issues in forensic pathology.^10,11 A recent systematic review involving 664 stillbirths highlighted a lack of validated, standardized methods to estimate the time of death in stillborn fetuses. ¹⁰ While postmortem imaging has been proposed for such estimation,^10,12-14 traditional evaluations remain predominant. These include the macroscopic assessment of the extent and location of fetal maceration, as well as the histological evaluation of autolysis in fetal tissues (loss of nuclear basophilia),^15-17 that progresses from renal cortical tubes to the gastrointestinal tract, liver, myocardium, and other organs, and in placental tissues.^5,18

Maceration refers to the degenerative changes arising in skin and tissues following the intrauterine death, due to a softening effect of soaking in the amniotic fluid.^17-19 No changes or little modifications are typically described in fetuses who remain in utero less than 8 hours, followed by a gradual peeling of the skin (grade 1 maceration according to Langley or “slight maceration according to Bain). With prolonged time in utero, the peeling progresses and becomes widespread, coupled to abdomen discoloration (1-2 days), diffuse reddening, uniform coloration of organs, red effusions in the serous cavities (grade 2 maceration according to Langley, corresponding to 2-3 days or “moderate maceration for Bain, corresponding to 2-7 days). Later on, sero-sanguinous effusions, overlapping skull bones, sunken eyes (4-7 days) as well as brown discoloration (>7 days) are described (grade III maceration according to Langley and “severe” maceration according to Bain).^5,19,20 Hill et al ²¹ described the loss of integrity of fetal epidermis within 3 to 4 days after death, followed by subcutaneous edema at ultrasound, skull bones overlap, and protrusion of brain tissue into the subcutaneous tissues of the neck.

Genest and Singer better detailed a discoloration of the umbilical cord and a skin slippage of at least 1 cm 6 hours after death, then peeling involving face, abdomen and back after at least 12 hours, desquamation involving 5% of body surface after 18 hours, and brown discoloration with extensive skin peeling after 24 hours.^17,19

Despite its wide use, the macroscopic evaluation of maceration is limited by several factors, including subjectivity, variability in reported timings across publications, poor consideration of the delivery-to-autopsy time, and poor interobserver reproducibility.^10,19,22

Interestingly, maceration-like skin changes are not uniquely encountered in the postmortem examination of stillbirths. In the forensic investigations of drowning and submersion deaths, similar postmortem changes are a key component of the Total Aquatic Decomposition Score (TADS), used to estimate the postmortem submersion interval (PMSI).^23-25 This approach offers a precedent for applying a quantitative scoring system to assess time-dependent macroscopic external postmortem changes.

The present study, drawing an analogy with aquatic decomposition scores, aims to propose a quantitative scoring system for evaluating the macroscopic maceration process in stillbirths, based on retrospective photographic analysis. The score’s predictive value for estimating the postmortem interval was assessed by statistical modeling, cross-validation, and inter-rater agreement.

Materials and Methods

Measuring Decomposition and Development of an Atlas

To obtain a standardized method to quantify the degree of decomposition, a maceration score (MS) was developed, taking into consideration maceration trends already described in the literature^17-21 and considering only those characteristics that can be visually assessed from color photographs.

In analogy with aquatic scoring methods (TADS), to account for variation in the rate of decay among areas of the body, stages of maceration were identified by subscores for 4 areas: the head, the trunk, the limbs, and the cord. The subscores were incorporated into timetables for facial (FMS), body (BMS), limb (LMS), and cordonal MS (CMS; Table 1). The sum of the 4 subscores, the total macroscopic maceration score or TMMS, represents a quantitative stage of decomposition for the whole fetus body and ranges from a minimum of 4 to a maximum of 20.

OPEN IN VIEWER

Table 1.

Descriptive Stages for Maceration as Observed in Face (FMS), Body (BMS), Limbs (LMS), and Cord (CMS).

Score	FMS	BMS	LMS	CMS
1	No visible changes or	No visible changes	No visible changes	No visible changes
	Only palpebral edema
2	Red skin discoloration or	Red skin discoloration or	Red skin discoloration or	Red discoloration of umbilical cord stump
	Bulla formation or early skin slippage (<5% of the face and neck)	Bulla formation or early skin slippage (<5% of the trunk)	Bulla formation or early skin slippage (<5% of the trunk)
3	Initial skin slippage (<50% of the face and neck) or	Initial skin slippage (<50% of the trunk) or	Initial skin slippage (<50% of the trunk) or	Brown discoloration of umbilical cord stump
	Brown skin discoloration	Brown skin discoloration	Brown skin discoloration
4	Advanced skin slippage (>50% of the face and neck)	Advanced skin slippage (>50% of the trunk) or	Advanced skin slippage (>50% of the trunk) or	Gray-green discoloration of umbilical cord stump
		Abdominal bloating	Joint visible laxity/deformity
5	Skull bones overlapping or	Gray-green skin discoloration	Gray-green skin discoloration
	Sunken eyes or
	Lowered ears
6	Gray-green skin discoloration

A reference atlas, containing 45 color photographs corresponding each to one of the described stages of the TMMS, was developed by a 10-year-experienced forensic pathologist (A.G.) and a 25-year-experienced pathologist in fetal deaths (M.P.B) as a visual resource and example for scoring for additional raters. The color photographs used for the atlas were selected from SB not included in the present casuistry.

The atlas developed was used to train additional groups or individual raters, beside the above-mentioned group of experienced pathologists (Group-rater 1): a pair of residents in forensic pathology with 2 years of experience in forensic pathology, who worked collaboratively and solved discrepancies by referencing to the atlas (Group-rater 2), and a neonatologist (Rater 3). Each rater or group of raters was blinded to the TMMS scores assigned by the others.

All 52 included SB cases had available color photographs from postmortem examination that were used to measure decomposition (TMMS assignment and TMMS -PMI association study). Additionally, 2 cases had delivery photographs available, which were used solely for TMMS assignment and confrontation between TMMS assigned at delivery and postmortem examination. Consequently, TMMS was assigned by Group-rater 1, 2, and Rater 3 (for inter-rater reliability assessment) on 54 photograph sets corresponding to 52 SB.

Results

TMMS Score

The degree of maceration within the study sample varied from the absence of maceration with a relatively “fresh” appearance, corresponding to a TMMS of 4, to extensive maceration, with a TMMS of 20, which was assigned in 1 case of fetal reduction of multiple pregnancy.

The score was assigned to the 54 sets of color photographs, including 52 sets of photographs taken at the postmortem examination (pPMI) and 2 sets taken at the time of delivery (dPMI). As shown in Figure 1, in these 2 cases the TMMS increased from 5 at dPMI to 12 at pPMI and from 7 dPMI to 12 pPMI.

OPEN IN VIEWER

Figure 1.

Comparison of TMMS score assigned on the basis of color photographs taken at delivery (A and C) and at postmortem examination (B and D), respectively. In A and B, the TMMS progressed from 5 to 12, while in C and D it was increased from 7 to 12.

Each score of the TMMS, from 4 to 20, was assigned at least once among our stillborn, except for the scores 17 and 19. The lower scores (TMMS = 4 and 5) were well represented, the score 10 was the most frequent (14.9% of cases), while the upper scores (TMMS = 18 and 20) were only assigned once (Figure 2).

OPEN IN VIEWER

Figure 2.

Graphical distribution of TMMS values assigned to the study population.

The TMMS score did not differ on the basis of the trimester (Figure 3(A)) and cause of death (in both cases P > .05; Figure 3(B)), while higher TMMS was detected in cases with subacute-chronic distress compared to acute-subacute ones (P = .0207; Figure 3(C)).

OPEN IN VIEWER

Figure 3.

Graphical representation of the TMMS assigned depending on the trimester (A), cause of death (B), and fetal distress (C).

Inter-Rater Reliability

Statistical analyses on TMMS assigned to 54 sets of photographs revealed a high level of correlation between more (Group rater 1) and less experienced raters (Group rater 2), with Spearman r = .9380, as well as between Group rater 1 and Rater 3, who had a clinical background (r = . 8869).

Accordingly with the results of the Spearmen test, the interrater weighed agreement was 92.96% between forensic pathologists with different experience (Group rater 1 vs Group rater 2, with kappa = .7818). The agreement was slightly lower between Group rater 1 and the neonatologist (Rater 3), showing an agreement of 87.57% (kappa = .6343). Nevertheless, all kappa values accounted for a substantial agreement. ²⁷

A graphical representation of the TMMS values assigned by the 3 independent raters to the 54 sets of photographs is shown in Figure 4. Consistency in color – where darker shades indicate lower scores and lighter shades indicate higher scores – reflects strong agreement among raters.

OPEN IN VIEWER

Figure 4.

Heatmap showing the TMMS assigned by 3 raters (rows: Group rater 1, Group rater 2 and Rater 3 to 54 set of color photographs taken at postmortem examination or at delivery (columns). Each cell’s color of the heatmap indicates the TMMS value, ranging from 4, in black color, to 20, in light yellow, as shown in the legend on the right. Consistent colors across a column indicate strong agreement among raters for that photographic set, while variation in colors within a column reflects inter-rater variability in scoring.

TMMS and PMI

The TMMS showed a significant correlation with the dPMI (P = .001) with good values (r = .8507, and 95% confidence interval from 0.7392 to 0.9168). Higher r values were obtained when associating TMMS with pPMI (P = .001): r = .9644 with 95% confidence interval from 0.9362 to 0.9803.

On this basis, pPMI was selected for the regression analyses.

Linear and multiple non-linear regressions were attempted between TMMS and pPMI. According to statistical analyses, the best fit-to-data was obtained, as shown in Figure 5, with a second order polynomial regression, which displayed a R² value of .958. This curve is characterized by 3 parameters: an intercept, a linear, and a quadratic term (shown in Table 2, together with 95% confidence intervals). With these curve parameters, the estimated pPMI for each TMMS score was calculated, together with their 95% confidence interval, as shown in Table 2.

OPEN IN VIEWER

Figure 5.

Second order polynomial regression curve linking pPMI, at the x axis, expressed as base 10 logarithm, to TMMS, at the y axis.

OPEN IN VIEWER

Table 2.

Results of the Parameters of the Second Order Polynomial Regression, With Best-fit Values and 95% Confidence Intervals (CI), and Predicted pPMI Values for Each TMMS Score, Also With Lower and Upper CI.

	Best-fit values	95% CI
Intercept (B0)	−1.1019	−2.1918, −0.0119
Linear term (B1)	27.9758	24.7498, 31.2017
Quadratic term (B2)	9.3332	−11.4717, −7.1946
TMMS	Predicted pPMI	Lower CI	Upper CI
4	1.57	1.48	1.65
5	1.73	1.64	1.81
6	1.91	1.82	1.98
7	2.11	2.03	2.19
8	2.35	2.27	2.43
9	2.63	2.54	2.72
10	2.96	2.86	3.06
11	3.34	3.22	3.47
12	3.81	3.66	3.98
13	4.38	4.18	4.60
14	5.08	4.82	5.39
15	5.98	5.61	6.43
16	7.17	6.63	7.83
17	8.81	7.99	9.92
18	11.27	9.87	13.72
19	15.66	12.56	21.66
20		16.22	30.81

Lastly, a cross-validation was performed, leading to a calculated mean absolute error of 0.39 days, root mean square error of 0.5991, and mean relative error of 9.40%. Absolute errors as calculated for TMMS intervals 4 to 7, 8 to 11, 12 to 15, and 16 to 19 are shown in Figure 6, and the Kruskal Wallis demonstrated significant differences (P = .0004).

OPEN IN VIEWER

Figure 6.

Absolute errors in days shown for TMMS values −7, 8 to 11, 12 to 15, and 16 to 19.

Discussion

The general sequence of macroscopic changes occurring in utero after death has already been well described, though the timing of these changes varies from one study to another.^17-21

An attempt toward a more objective evaluation of the maceration process was suggested by the Stillbirth Collaborative Research Network, which graded maceration on a scale from 0 to 5, ²⁸ but the grading method was not applied to the collected stillbirths and there is no data on the accuracy of this scoring system.

Given the importance of timing in stillbirths, especially in cases of suspected medical malpractice, the present study proposes a scoring method to quantify the degree of maceration, which has previously been reported in descriptive and qualitative terms.^10,22

Although the fetus is entirely immersed in amniotic fluid, the maceration process does not affect all areas of the body simultaneously, as suggested by Genest and Singer, ¹⁷ who first described a discoloration of the umbilical cord stump. Accordingly, varying rates of maceration in different body regions, were accounted for using subscores.

The process of maceration is considered to cease once the fetus is delivered. ²⁹ However, as noted with submerged bodies, once removed from the fluid, the body may undergo even accelerated putrefaction. ³⁰ The delivery-to-autopsy interval has received little consideration in previous literature regarding PMI in SB, and this represents a limitation of PMI estimation methods based on appearance of fetuses. ¹⁰

To overcome this, our study considered both dPMI (calculated at the time of delivery) and pPMI (calculated at postmortem examination) for their association with the TMMS in 47 stillbirths. As expected, a stronger association was found with the pPMI, since photographs were for the majority taken at postmortem examination. This result was further supported by comparing TMMS assigned to photographs collected at delivery vs at postmortem examination in 2 cases of our casuistry, resulting in a TMMS increase from the dPMI to the pPMI. Although the number of cases was very limited, this finding suggests that the delivery-to-postmortem examination period should be considered for PMI estimation, and that maceration might continue and overlap with putrefaction, especially when storage temperatures are cold but not freezing.

When examining the relationship between TMMS and pPMI, the trend appeared as a concave-down parabola, suggesting that TMMS scores increase with pPMI up to a certain point and then begin to plateau. This observation is consistent with the description of the maceration process found in the literature, which does not include specific changes after 1/2 weeks depending on the used score.^17-21

The R² (coefficient of determination) value of .958 for the second order polynomial curve indicates that approximately 95.8% of the variance in the dependent variable can be explained by the independent variables in our model. This suggests that the model fits the data quite well, and is higher that the corresponding values reported for the postmortem submersion interval (R² ranging from .77 to .91).^23,25 Despite the good results, it should be noted that some cases had to be excluded as potential outliers, and this was connected to the storage conditions of the bodies. This factor limits the possibility of applying the score when no data are available on the delivery-to-autopsy interval

The accuracy in predicting pPMI from the TMMS, evaluated through a cross-validation in the present study, demonstrated an absolute mean error of 0.39. It is rather challenging to compare this result with previous literature, due to varied methodologies used. Using the external fetal examination criteria reported by Genest, 69% of fetuses were correctly classified. ¹⁷ According to Gold et al, ²² the external appearance of fetuses, whether “fresh” or “macerated,” did not allow for a good estimation of the death-to-delivery interval or prepartum vs intrapartum demise. They reported an accuracy of 22% for this method. ¹⁰ To the best of our knowledge, this is the first score to allow an error estimation, and thus represent a step toward a probabilistic approach to the assessment of PMI in stillbirths.

The higher accuracy at lower TMMS scores in our study was expected, considered that external maceration features better reflect early rather than late changes. ¹⁷ Nevertheless, this limitation could be overcome by combining TMMS with histopathological examination of organs as well as with the imaging-based scores, which have shown high sensitivity for detecting severe maceration at autopsy.^12-14

In addition to a low absolute error, the inter-rater reliability results showed a substantial agreement between raters with varying levels of experience and background, with values similar to those reported for imaging-based maceration scores. ¹³

This result suggests that the TMMS score is a promising tool that could be applied in combination with current gold-standard methods to increase the accuracy of estimating the time of intrauterine deaths in stillbirths.

Moreover, since skin maceration is often described directly by birth attendants, including gynecologists, nurses, or midwives, ³¹ the substantial rater agreement found in our study suggests that TMMS could be easily and objectively applied, with further training and testing among additional healthcare professionals, even at the time of delivery. This approach would likely enhance the association between TMMS and dPMI, potentially surpassing the current association observed between TMMS and pPMI, by eliminating the delivery-to-autopsy interval, thereby reducing a source of uncertainty. Indeed, pathologists might not be aware of the specific storage conditions of the body, which can introduce variability in the maceration process.

It is well known that some factors might influence the maceration process, for example a high microbial load in the amniotic fluid, maternal fever, or a long duration of hypoxia prior to actual death.^19,22 In our casuistry, TMMS was not significantly higher in cases with fatal infection compared to other causes of death. On the other hand, an influence of subacute-chronic distress was shown, confirming that prolonged hypoxia might lead to accelerated maceration and higher scores. The accuracy of predicting the PMI from TMMS would benefit from further studies assessing the impact of fetal distress.

The present study has several limitations, with the major drawback being the uncertainty regarding the exact timing of intrauterine fetal death. To address it, a multidisciplinary evaluation of all available data (including circumstantial and clinical information, as well as postmortem examination with macroscopic assessment and histopathology of fetal tissues and annexes) was employed. This approach is currently regarded as the gold-standard for achieving an estimate as close as possible to the true value. Although most studies on this topic are retrospective,^10,12,17 future research could be designed either prospectively, in order to further reduce the issue of data heterogeneity, or including only cases with known time of death, though this might hamper the collection of a large sample.

The ability to score maceration from photographs was clearly dependent on the quantity and quality of the photographs, as for TADS in submerged bodies, ²³ but raters did not report any instances of missing, insufficient data, or low resolution photographs. Termination of multiple pregnancies were included, because they provided the upper limit of the score and represented cases of known timing of death.

A strength of the study lies in the collection of extensive data on each included case, which likely contributed to a limited sample size. The consideration of delivery-to-autopsy timing, cross-validation, and assessment of inter-rater reliability represent further strengths of our study, and contributed to the promising results for estimating PMI in stillbirths.

Although further research would benefit from a larger study sample and the involvement of different specialties, the here-proposed score might be used to standardize data collection, with the aim of improving the ability to assess the time of death, as well as to improve the current practice in perinatal death surveillance.

While Genest scoring remains the standard for estimating PMI in stillbirths, our TMMS scoring approach offers potentially improved precision as indicated by narrower estimated time intervals and higher inter-rater reliability, and provides an estimation of error, useful in courtroom.

The present proposal was a first, preliminary step toward better assessment of PMI in stillbirths, although the score was exclusively based on gross fetal features, and a direct comparison between the 2 scoring systems was beyond the scope of the study. A more comprehensive and integrative approach, combining TMMS with fetal and placental histological findings and/or integrating imaging, will be essential to surpass Genest-based assessments and to obtain a score reliable from a forensic and diagnostic point of view.

Conclusions

The present preliminary study proposes a scoring method, the total macroscopic maceration score or TMMS, to assess the grade of maceration of stillbirths, which could be easily applied by forensic pathologists as well as by clinicians attending birth. Integrated among the current gold-standard, with a comprehensive approach also encompassing fetal and placental histological findings and coupled to imaging-based methods, the TMMS might represent an additional tool to assess with accuracy the timing of intrauterine fetal deaths.