Advertisement

European Radiology

, Volume 26, Issue 4, pp 1186–1190 | Cite as

Distinction between saltwater drowning and freshwater drowning by assessment of sinus fluid on post-mortem computed tomography

  • Yusuke KawasumiEmail author
  • Akihito Usui
  • Yuki Sato
  • Yumi Sato
  • Nami Daigaku
  • Yoshiyuki Hosokai
  • Yoshie Hayashizaki
  • Masato Funayama
  • Tadashi Ishibashi
Forensic Medicine

Abstract

Objective

To evaluate the difference in sinus fluid volume and density between saltwater and freshwater drowning and diagnose saltwater drowning in distinction from freshwater drowning.

Methods

Ninety-three drowning cases (22 saltwater and 71 freshwater) were retrospectively investigated; all had undergone post-mortem CT and forensic autopsy. Sinus fluid volume and density were calculated using a 3D-DICOM workstation, and differences were evaluated. Diagnostic performance of these indicators for saltwater drowning was evaluated using a cut-off value calculated by receiver operating characteristic (ROC) analysis.

Results

The median sinus fluid volume was 5.68 mL in cases of saltwater drowning (range 0.08 to 37.55) and 5.46 mL in cases of freshwater drowning (0.02 to 27.68), and the average densities were 47.28 (14.26 to 75.98) HU and 32.56 (−14.38 to 77.43) HU, respectively. While sinus volume did not differ significantly (p = 0.6000), sinus density was significantly higher in saltwater than freshwater drowning cases (p = 0.0002). ROC analysis for diagnosis of saltwater drowning determined the cut-off value as 37.77 HU, with a sensitivity of 77 %, specificity of 72 %, PPV of 46 % and NPV of 91 %.

Conclusion

The average density of sinus fluid in cases of saltwater drowning was significantly higher than in freshwater drowning cases; there was no significant difference in the sinus fluid volume.

Key points

Sinus fluid density of saltwater drowning is significantly higher than freshwater drowning.

Cut-off value was 37.77 HU based on the ROC analysis.

The cut-off value translated to 91 % NPV for diagnosis of saltwater drowning.

Keywords

Drowning Freshwater Saltwater Paranasal sinuses Computed tomography 

Abbreviations

CT

Computed tomography

MDCT

Multi-detector computed tomography

DICOM

Digital imaging and communication in medicine

ROC

Receiver operating characteristic

PPV

Positive predictive value

NPV

Negative predictive value

Introduction

When a dead body is found in or around water, drowning must be considered as a cause of death. Common macroscopic findings of drowning are froth in the airway, waterlogged and over-distended lungs, pleural effusion, fluid collection in the stomach, and reduced spleen mass, among others [1, 2, 3]. However, these findings are absent in some drowning cases because the body is found long after death or has decomposed. In these cases, microscopic findings such as diatoms in the lungs, spleen, kidneys, and bone marrow serve as a basis for diagnosis of drowning [1, 2, 3]. Moreover, forensic pathologists need to specify where and why drowning occurred. Evidence of diatoms in the submerging fluid and in the victim’s body represents the current gold standard for diagnosing type of drowning [4]. Biochemical and immunohistochemical approaches for differential diagnosis between saltwater and freshwater drowning have also been described [5, 6, 7].

In forensic medicine, post-mortem computed tomography (CT) has become common [8, 9, 10, 11]. Various CT findings of drowning have been reported, and one of the most notable findings is fluid accumulation in the maxillary or sphenoid sinuses [12, 13]. Lo Re G. et al. [14] also diagnosed drowning when fluids were present in the paranasal sinuses of drowned bodies. To differentiate non-drowned and drowned cadavers, an evaluation of differences in the density of the blood vessels (the resorption of freshwater in the lung results in hypodensity of the blood, representing haemodilution), and numerous other signs of drowning, including the density of the airway contents in HU, have been reported [12, 15]. Previously, we found that sinus fluid volume is significantly greater and sinus fluid density is significantly lower in drowning cases than in non-drowning cases [16]. Although these findings can help forensic pathologists distinguish between drowning and non-drowning, they cannot provide information concerning the scene of drowning. No reported studies have suggested that differences in the density of sinus fluid content could be used to differentiate between freshwater and saltwater drowning. Therefore, in this study, we examined differences on post-mortem CT between saltwater and freshwater drowning cases.

Materials and methods

Study cases

We retrospectively reviewed 641 cases that had undergone post-mortem CT and a forensic autopsy between May 2009 and January 2013 at our institution. There were 119 drowning cases. We excluded 3 infant cases and 10 cases with pronounced decomposition, and finally selected 106 drowning cases. All were diagnosed as wet drowning by autopsy and had fluid accumulation in the maxillary or sphenoid sinuses on post-mortem CT. The autopsies also confirmed that there were no head or facial fractures causing haemosinus. Typical CT images of fluid accumulation are shown in Fig. 1. Study cases included 22 saltwater drownings (10 females, 12 males; mean age 60.3 ± 13.4 (range 35–87) years) and 71 freshwater drownings (21 females, 50 males; mean age 64.6 ± 17.4 (range 19–100) years). Saltwater drowning cases occurred at sea only; there were no saltwater lake cases. The scenes of freshwater drowning were river (n = 28), irrigation channel (n = 16), street gully (n = 6), rice paddy (n = 4), dam (n = 3), pond (n = 3), sewer (n = 2), well (n = 2), and others (n = 7). The median interval between death and CT was 1 day (range 0.5 to 10 days). This retrospective study was approved by the Ethics Board of our institution. Informed consent was not required, as all image data were obtained previously.
Fig. 1

Post-mortem cranial CT of typical drowning cases showing fluid accumulation in the maxillary sinus (arrows). (a) A case of saltwater drowning (volume 4.68 mL, average density 46.46 HU). (b) A case of freshwater drowning (volume 4.23 mL, average density 13.44 HU)

CT and autopsy

An eight-channel multi-detector row CT (MDCT) scanner (Aquilion 8; Toshiba Medical Systems, Tokyo, Japan) was used for all examinations. All subjects were scanned in a body bag while fully clothed. No contrast material was administered. In all cases, the CT scan was taken from the head to the pelvis in the helical mode (tube voltage, 120 kVp; installed mode tube current; rotation time, 0.75 sec; table speed, 14 mm per rotation; helical beam pitch, 0.875; collimations, 2.0 mm; reconstruction interval, 0).

In all cases, a forensic autopsy was performed shortly after the CT scan by a forensic pathologist with more than 30 years of experience.

Image assessment

All 2.0-mm slice CT image data were sent to a three-dimensional DICOM workstation (Ziostation version 2.0.0.1; Ziosoft, Tokyo, Japan). Areas of fluid in the maxillary and sphenoid sinuses on axial images were extracted manually for each slice. No multiplanar reconstruction (MPR) images were used during this process. The workstation automatically calculated the volume data and average density of the extracted fluid areas during three-dimensional volume rendering reconstruction.

Statistical analysis

Normality tests indicated that the fluid volume was not, but the fluid density was, normally distributed. Therefore, statistical significance was evaluated by the Mann–Whitney U-test and Student’s t-test, respectively. A p-value of less than 0.05 was considered to indicate a significant difference. Receiver operating characteristic (ROC) analysis was performed to determine the cut-off level, from the point on the ROC curve closest to (0, 1); i.e. the point nearest to the top left corner of the graph, and the Youden index. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for a diagnosis of saltwater drowning were calculated based on this cut-off value.

Results

Figure 2 shows a Tukey boxplot of the volume distribution of fluid accumulation in the maxillary or sphenoid sinuses. In saltwater drowning cases, the median volume was 5.68 (range 0.08–37.55) mL, versus 5.46 (0.02–27.68) mL in freshwater drowning cases. Fluid volume did not significantly differ between the saltwater and freshwater drowning cases (p = 0.6000, Mann–Whitney U-test).
Fig. 2

Tukey boxplot of volumes of accumulated fluid in the maxillary and sphenoid sinuses. The median value was 5.68 mL in saltwater drowning cases and 5.46 mL in freshwater drowning cases. The difference was not significant (p = 0.6000, Mann–Whitney U-test)

Figure 3 shows a Tukey boxplot of the density distribution of fluid accumulation in the maxillary or sphenoid sinuses. In saltwater drowning cases, the average density was significantly greater than in fresh water drowning cases (47.28 (range 14.26 to 75.98) HU versus 32.56 (−14.38 to 77.43) HU; p = 0.0002; Student’s t-test). When the point on the ROC curve (Fig. 4) closest to (0, 1) was set as the cut-off level, the cut-off value was 37.77 HU, sensitivity 77 %, specificity 72 %, PPV 46 %, and NPV 91 %. The Youden index was the same as the point on the ROC curve closest to (0, 1).
Fig. 3

Tukey boxplot of densities of accumulated fluid in the maxillary and sphenoid sinuses. The average value was 47.28 HU in saltwater drowning cases and 32.56 HU in freshwater drowning cases. Sinus fluid density in saltwater drowning cases was significantly higher than that in freshwater drowning cases (p = 0.0002; Student’s t-test)

Fig. 4

ROC curve for diagnosis of saltwater drowning based on maxillary and sphenoid sinus fluid density. Both the point on the ROC curve closest to (0, 1) and the Youden index indicated a cut-off value of 37.77 HU, yielding a sensitivity of 77 %, specificity 72 %, PPV 46 %, and NPV 91 %

Discussion

When investigating drowning cases, the scene of drowning must be determined. Histopathological, biochemical, and immunohistochemical approaches have been described [4, 5, 6, 7, 17, 18, 19, 20], but the feasibility of drowning site determination by post-mortem CT has not been investigated. Here, we demonstrate that the density of fluid accumulated in the maxillary and sphenoid sinuses was significantly higher in saltwater drowning cases than in freshwater drownings.

In our previous study, we determined that sinus fluid volume following drowning was significantly greater than in cases that had died of other causes [21]. While the volume distribution might be presumed to depend on the composition of fluid among drowning cases, this study established that there was no significant difference in the volume between saltwater and freshwater drowning. Thus, it is not possible to presume the scene of drowning based on the fluid volume.

As the density of saltwater is higher than that of freshwater, CT attenuation values for sinus fluids following drowning would be expected to reflect this difference. Nonetheless, the validity of CT image analysis for distinguishing these fluids had not been investigated in drowning victims. This study has revealed that maxillary and sphenoid sinus fluid density as determined by post-mortem CT in saltwater drowning cases is higher than following freshwater drowning. Usumoto et al. [22] reported that the pleural effusion resulting from seawater drowning was characterised by a higher concentration of sodium and chloride ions compared to freshwater drowning. These differences are presumed to also be present in the paranasal sinuses and can probably partially explain the differences in fluid density. In the ROC analysis for diagnosis of saltwater drowning, the candidate cut-off value was 37.77 HU, which translated to 77 % sensitivity, 72 % specificity, 46 % PPV, and 91 % NPV. Therefore, saltwater drowning is highly unlikely when the sinus fluid density of a drowning victim is less than 37.77 HU on post-mortem CT. If a victim found in the sea or coast has sinus fluid less dense than 37.77 HU, the possibility that the victim may have drowned at another place must be considered. We also evaluated the average density of the extracted fluid areas, but did not consider differences in the density of regions within the same area. When differences were large, evaluating only the average density could constitute a methodological bias.

In our study, all diagnostic indicators of saltwater drowning were inadequate except for NPV. Christe et al. [12] reported that the froth in the airways, emphysema aquosum and mosaic patterns in the lung on CT were all suggestive of freshwater drowning. Ambrosetti et al. [15] reported that the measurement and comparison of blood-density values inside cardiac chambers may be reliable because any differences should be caused by differences in the pathophysiological mechanism underlying freshwater versus saltwater drowning; using these data could increase diagnostic performance.

Cases with pronounced decomposition were excluded from our study because saprogenous exudate would be miscible with water and affect sinus fluid density. Infant cases were also excluded because their paranasal sinuses are still unformed. As a result of the exclusion, the youngest case in this study was 19 years old; his sinuses could be assumed to be mature.

The freshwater drowning cases in this study drowned at various scenes. Differences in drowning location may be associated with differences in sinus fluid volume and density, even among freshwater drowning cases. However, we did not examine the differences in fluid volume and density among these locations because the number of drownings at each type of location was too small to be assessed statistically. This is thus a subject for future investigation.

Reproducibility of the volume and density values was not perfect because areas of fluid in the maxillary and sphenoid sinuses were extracted manually. Additionally, the volume and density values were not compared with values collected at autopsy because these data were not available for all cases.

Conclusion

The average density of sinus fluid in cases of saltwater drowning was significantly higher than in freshwater drowning cases, while there was no significant difference in the sinus fluid volume.

Notes

Acknowledgments

The scientific guarantor of this publication is Prof. Tadashi Ishibashi. The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article. The authors state that this work has not received any funding. No complex statistical methods were necessary for this paper. Institutional Review Board approval was obtained. Written informed consent was obtained from all subjects (patients) in this study. Methodology: Retrospective, diagnostic or prognostic study, performed at one institution.

References

  1. 1.
    Dolinak D, Matshes E, Lew E (eds) (2005) Forensic pathology principles and practice. Elsevier, AmsterdamGoogle Scholar
  2. 2.
    Tsokos M (ed) (2005) Forensic pathology reviews volume 3. Humana Press, New YorkGoogle Scholar
  3. 3.
    Saukko P, Knight B (eds) (2004) Knight’s forensic pathology, 3rd edn. Edward Arnold, LondonGoogle Scholar
  4. 4.
    Pollanen MS, Cheung C, Chiasson DA (1997) The diagnostic value of the diatom test for drowning, I. utility: a retrospective analysis of 771 cases of drowning in Ontario, Canada. J Forensic Sci 42:281–285PubMedGoogle Scholar
  5. 5.
    An JL, Ishida Y, Kimura A, Kondo T (2010) Forensic application of intrarenal aquaporin-2 expression for differential diagnosis between freshwater and saltwater drowning. Int J Legal Med 124(2):99–104CrossRefPubMedGoogle Scholar
  6. 6.
    An JL, Ishida Y, Kimura A, Kondo T (2011) Immunohistochemical examination of intracerebral aquaporin-4 expression and its application for differential diagnosis between freshwater and saltwater drowning. Int J Legal Med 125(1):59–65CrossRefPubMedGoogle Scholar
  7. 7.
    Cala AD, Vilain R, Tse R (2013) Elevated postmortem vitreous sodium and chloride levels distinguish saltwater drowning (SWD) deaths from immersion deaths not related to drowning but recovered from saltwater (DNRD). Am J Forensic Med Pathol 34(2):133–138CrossRefPubMedGoogle Scholar
  8. 8.
    Thali MJ, Braun M, Buck U et al (2005) VIRTOPSY-scientific documentation, reconstruction and animation in forensic: individual and real 3D data based geo-metric approach including optical body/object surface and radiological CT/MRI scanning. J Forensic Sci 50:428–442CrossRefPubMedGoogle Scholar
  9. 9.
    Dirnhofer R, Fackowski C, Vock P, Potter K, Thali MJ (2006) VIRTOPSY: minimally invasive, imaging-guided virtual autopsy. Radiographics 26:1305–1333CrossRefPubMedGoogle Scholar
  10. 10.
    Bolliger SA, Thali MJ, Ross S, Buck U, Naether S, Vock P (2008) Virtual autopsy using imaging: bridging radiologic and forensic sciences. a review of the virtopsy and similar projects. Eur Radiol 18:273–282CrossRefPubMedGoogle Scholar
  11. 11.
    Le Blanc-Louvry I, Thureau S, Duval C et al (2013) Post-mortem computed tomography compared to forensic autopsy findings: a French experience. Eur Radiol 23(7):1829–1835CrossRefPubMedGoogle Scholar
  12. 12.
    Christe A, Aghayev E, Jackowski C, Thali MJ, Vock P (2008) Drowning—post-mortem imaging findings by computed tomography. Eur Radiol 18:283–290CrossRefPubMedGoogle Scholar
  13. 13.
    Levy AD, Harcke HT, Getz JM et al (2007) Virtual autopsy: two- and three-dimensional multidetector CT findings in drowning with autopsy comparison. Radiology 243(3):862–868CrossRefPubMedGoogle Scholar
  14. 14.
    Lo Re G, Vernuccio F, Galfano MC et al (2015) Role of virtopsy in the post-mortem diagnosis of drowning. Radiol Med 120(3):304–308CrossRefPubMedGoogle Scholar
  15. 15.
    Ambrosetti MC, Barbiani C, El-Dalati G et al (2013) Virtual autopsy using multislice computed tomography in forensic medical diagnosis of drowning. Radiol Med 118:679–687CrossRefPubMedGoogle Scholar
  16. 16.
    Kawasumi Y, Kawabata T, Sugai Y et al (2013) Diagnosis of drowning using post-mortem computed tomography based on the volume and density of fluid accumulation in the maxillary and sphenoid sinuses. Eur J Radiol 82(10):e562–e566CrossRefPubMedGoogle Scholar
  17. 17.
    Byard RW, Cains G, Tsokos M (2006) Haemolytic staining of the intima of the aortic root-- a useful pathological marker of freshwater drowning? J Clin Forensic Med 13(3):125–128CrossRefPubMedGoogle Scholar
  18. 18.
    Azparren JE, Cubero C, Perucha E, Martinez P, Vallejo G (2007) Comparison between lung weight and blood strontium in bodies found in seawater. Forensic Sci Int 168(2–3):128–132CrossRefPubMedGoogle Scholar
  19. 19.
    Lucci A, Campobasso CP, Cirnelli A, Lorenzini G (2008) A promising microbiological test for the diagnosis of drowning. Forensic Sci Int 182(1–3):20–26CrossRefPubMedGoogle Scholar
  20. 20.
    Maeda H, Zhu BL, Ishikawa T et al (2009) Analysis of postmortem biochemical findings with regard to the lung weight in drowning. Leg Med 11(Suppl 1):S269–S272CrossRefGoogle Scholar
  21. 21.
    Kawasumi Y, Kawabata T, Sugai Y et al (2012) Assessment of the relationship between drowning and fluid accumulation in the paranasal sinuses on post-mortem computed tomography. Eur J Radiol 81(12):3953–3955CrossRefPubMedGoogle Scholar
  22. 22.
    Usumoto Y, Sameshima N, Hikiji W et al (2009) Electrolyte analysis of pleural effusion as an indicator of drowning in seawater and freshwater. J Forensic Leg Med 16:321–324CrossRefPubMedGoogle Scholar

Copyright information

© European Society of Radiology 2015

Authors and Affiliations

  • Yusuke Kawasumi
    • 1
    Email author
  • Akihito Usui
    • 2
  • Yuki Sato
    • 1
  • Yumi Sato
    • 1
  • Nami Daigaku
    • 2
  • Yoshiyuki Hosokai
    • 2
  • Yoshie Hayashizaki
    • 3
  • Masato Funayama
    • 3
  • Tadashi Ishibashi
    • 1
  1. 1.Department of Clinical ImagingTohoku University Graduate School of MedicineSendaiJapan
  2. 2.Department of Diagnostic Image AnalysisTohoku University Graduate School of MedicineSendaiJapan
  3. 3.Department of Forensic MedicineTohoku University Graduate School of MedicineSendaiJapan

Personalised recommendations