Study of the factors at admission predicting the outcome in patients with attempted suicidal hanging

Abstract

Our study sought to identify factors at presentation which can predict the outcome after an attempted hanging. A retrospective analysis of patients over a 12-year period was carried out. A poor outcome was found in 17.8% and this could be predicted by the presence of myoclonus, a Glasgow coma motor score of ≤3 or an abnormal chest radiograph.

Keywords

Hanging outcome myoclonus Glasgow Coma Score (GCS)motor score of GCS

Introduction

The rate of adult suicide in Tamil Nadu, the state where our hospital is situated, is more than double the national average.¹ The National database shows that hanging is the most common method of suicide with an increase in incidence from 37% in 2012 to 41.8% in 2014.¹

Published studies on hanging generally describe small cohorts of patients with an emphasis on factors predicting mortality.^2–4 There is a gap in the available literature concerning factors which affect the morbidity of near-hanging. In a situation of limited resources, we believe it is of prime importance to be able to predict morbidity and so better to allocate available resources.

Materials and methods

This was a retrospective observational study conducted in a tertiary care hospital in southern India over a 12-year period (January 2003 to January 2015) with permission from the Internal Review Board (IRB no. 10241 dated 24 August 2016). Clinical and laboratory data were collected by the investigators from computerised case records and the laboratory database. Data collected included duration of hospital stay, material used to for hanging, cause, history of previous attempts and psychiatric illness. Admission data gathered were heart rate, blood pressure (a systolic pressure < 90 mmHg being defined as hypotension), Glasgow Coma Score (GCS) and, in particular, the motor score (a good outcome being defined by a score of 4–6), pupillary size and reaction, presence of myoclonus, airway and breathing status. Cervical spine and chest radiographs were reviewed on our archival and communication system (PACS GE). The need and duration of ventilation, tracheostomy and the outcome at hospital discharge was documented. No follow-up after discharge was carried out.

Statistical analysis

Data were analysed using the Statistical Package for Social Services (SPSS) software Version 21.0 (IBM Corp., Armonk, NY, USA). Bivariate analysis was done using the Pearson chi-square test, Fisher’s exact test and non-parametric Mann–Whitney U test. Multivariate analysis was done for variables which had statistical significance on bivariate analysis.

Results

A total of 441 patients identified as cases of near-hanging were admitted in the 12-year period under the Department of Neurosciences. Of these, 32 patients were excluded due to unavailability of charts (n = 30) and incorrectly documented diagnosis (n = 2).

Of the remaining 409 patients, there was a slight female preponderance (n = 209, 51.1%) and the majority (n = 370, 90.5%) were aged ≤45 years (Table 1). A poor outcome at discharge (GOS 1–3, Table 2) was noted in 73 (17.8%), of whom 35 (8.6%) were dependent on others for activities of daily living at discharge, 32 (7.8%) died and six (1.5%) were discharged at the request of relatives (Table 2). The patients who were discharged at request were due to the irremediable condition of the patient (persistent GCS < 8 or ventilatory dependence).

Table 1.

Sample characteristics of the study population (n = 409).

Characteristics	Good* (n = 336)	Poor* (n = 73)	P value for outcomes
Gender			0.47
Women (n = 209, 51.1%)	164	45
Men (n = 200, 48.9%)	172	28
Age (years)			0.648
≤45 (n = 370, 90.5%)	305	65
>45 (n = 39, 9.5%)	31	8
Previous attempt			0.151
No previous attempt (n = 375, 91.7%)	305	70
Previous attempt (n = 34, 8.3%)	31	3
Psychiatric illness			0.241
Known (n = 37, 9%)	33	4
Not known (n = 372, 91%)	303	69
Material used†			0.069
Saree/cloth (n = 174, 68%)	153	21
Rope (n = 82, 32%)	65	17
Myoclonus			<0.0001
Absent (n = 355, 95.9%)	309	46
Present (n = 15, 4.1%)	6	9
Hypotension (systolic blood pressure < 90 mmHg)			0.882
Present (n = 81, 19.8%)	67	14
Absent (n = 328, 80.2%)	269	59
GCS–Motor score‡			<0.0001
1–3 (n = 153, 44.3%)	110	43
4–6 (n = 192, 55.7%)	177	15
GCS§			0.003
3–8 (n = 269, 68.8%)	214	55
≥9 (n = 122, 31.2%)	112	10
Ventilatory support			<0.0001
Ventilated (n = 347, 84.8%)	274	73
Not ventilated (n = 62, 15.2%)	62	0
Chest X-ray**			<0.0001
Normal (n = 202, 49.38%)	185	17
Abnormal (n = 196, 47.92%)	144	52
Duration of hospital stay (days)			<0.0001
0–5 (n = 257, 62.8%)	229	28
≥6 (n = 152, 37.2%)	107	45

GOS (Glasgow outcome scale) of 1–3 is a poor outcome and 4–5 is a good outcome.

†

Material used for hanging was not known in 153 cases.

‡

GCS–Motor score not documented in 64 cases.

GCS–Glasgow coma scale not documented in 18 cases.

Not available in 11 cases.

Table 2.

Outcome data.

Outcome	n = 409 (%)
Glasgow outcome scale 4–5	336 (82.2)
Glasgow outcome scale 2–3	41 (10.02)
Dead (Glasgow outcome scale 1)	32 (7.8)

Myoclonus at admission was seen in 15 patients (3.6%). Simple logistic regression analysis (Table 3) showed this to be significant in predicting a poor outcome (odds ratio [OR] = 10.1, 95% confidence interval [CI] = 3.4–29.6, P value < 0.0001). The GCS at presentation was in the range of 3–8 in 269 (65.7%) and was seen to be significant in affecting the outcome (OR = 2.9, 95% CI = 1.4–5.9, P value = 0.004). However, a GCS motor score of ≤3 was seen to have a much more significant impact on outcome (OR = 4.6, 95% CI = 2.4–8.7, P value < 0.0001).

Table 3.

Bivariate logistic regression analysis for variables that predicted a poor outcome.

Variable	OR (95% CI)	P value
Myoclonus (if present)	10.1 (3.4–29.6)	<0.0001
GCS Motor score (1–3)	4.6 (2.4–8.7)	<0.0001
Abnormal chest X-ray (at presentation)	3.9 (2.2–7.1)	<0.0001
GCS (3–8)	2.9 (1.4–5.9)	0.004

OR, odds ratio; CI, confidence interval.

A large majority, 347 (84.8%), required ventilation at admission (type 1 respiratory failure or GCS <8 needing airway protection). Whereas, the 62 (15.2%) patients who did not need ventilatory support had a good outcome. Hypotension was recorded in 81 (19.8%) and did not statistically affect the outcome (OR = 0.9, 95% CI = 0.5–1.8, P value = 0.88). A tracheostomy was performed in 26 (6.4%) patients.

The chest radiograph was abnormal (features of post obstructive pulmonary oedema or aspiration pneumonia) in 196 patients (47.9%) at admission, significantly affecting the outcome (OR = 3.9, 95% CI = 2.2–7.1, P value < 0.001).

None of the patients had cervical fractures. Cervical cord oedema at the level of C4–C5 was seen in one patient. He stayed in hospital for 24 days and had minimal neurological deficit at discharge. One patient had injuries to the hyoid and thyrohyoid membrane leading to emphysema and pneumothorax. Local injuries were seen in 11 patients (3.1%), the commonest being a rotator cuff injury.

Hospital stay was ≥6 days in 152 (37.2%). Analysis using the Kolmogorov–Smirnov test showed there was no statistically significant difference in the duration of stay between gender and outcome.

Multivariate logistic regression analysis (Table 4) showed that myoclonus, low motor component of the GCS score and an abnormal chest radiograph all had a major impact on the outcome, with the statistically most significant factor being the presence of myoclonus.

Table 4.

Association between risk factors on outcome using multivariate logistic regression analysis.

	Model 1*		Model 2†		Model 3‡		Model 4§
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Myoclonus	11.3 (3.0–42.6)	<0.0001	12.2 (3.8–39.4)	<0.0001	7.8 (2.2–28.2)	0.002	–	–
Abnormal chest X-ray	4.1 (1.9–8.7)	0.001	4.6 (2.5–8.7)	<0.0001	–	–	3.4 (1.7–6.8)	0.001
Motor score (1–3)	3.2 (1.6–6.2)	<0.0001	–	–	4.1 (2.2–7.9)	<0.0001	3.6 (1.9–6.9)	<0.0001

Model 1 includes myoclonus, abnormal chest X-ray and Motor score.

†

Model 2 includes myoclonus and abnormal chest X-ray.

‡

Model 3 includes myoclonus and GCS Motor score.

Model 4 includes abnormal chest X-ray and GCS Motor score.

OR, odds ratio; CI, confidence interval.

Other data, such as place, type, material used, precipitating factors for hanging and time to reaching medical care, were not universally available and hence could not be usefully analysed.

Discussion

Hanging is a commonly used means of attempting suicide in our environment. The term, near-hanging, refers to patients who survive a hanging injury long enough to reach medical attention.²

The influence of the GCS on outcome is inconsistent in the reported literature.^3,5,6 We found that a GCS < 8 was a predictor of poor outcome but a low GCS was also seen in 52% of patients with a good outcome (Table 1). The motor score component of the GCS, which has not been previously investigated as an independent variable in near-hanging, appears to be a much better indicator of the outcome than the total GCS score, similar to studies in head injury.⁷

The presence of myoclonus is a surrogate marker of hypoxic neuronal damage and was noted to be highly predictive of a poor outcome. This corroborates a finding that anoxic changes on computed tomography (CT) scan were independently associated with poor outcome.⁶

An abnormal chest radiogram suggestive of pulmonary involvement with pulmonary oedema or aspiration pneumonia was also a significant predictor of poor outcome. Previous studies have also shown that the mortality increased from 5% to 34.2% in the presence of pulmonary involment.⁸ Using multivariate logistic regression analysis (Table 4), we found that the presence of myoclonus with an abnormal chest radiograph had the worst outcome (Model 2), followed by a combination of myoclonus and a low GCS motor score (Model 3).

In contrast to previous studies, the presence of hypotension at presentation in our cohort did not correlate with poor outcome.³

Limitations of the study

The retrospective design of this study has limited data availability. Hence, we could not evaluate important parameters, such as time to presentation, which are known to impact outcome.⁵ Another important limiting factor of this study is that the outcome assessment has been done only at discharge. This study was also a single-centre study which limits the generalisability of the study findings.

Conclusion

In a large cohort of 409 patients, 82% had a good outcome. The presence of myoclonus at presentation with an abnormal chest radiogram and low motor component of the GCS score predict a poor outcome.

Footnotes

Acknowledgements

The authors thank Dr Vivek Mathew, Department of Neurology, for his guidance and encouragement to complete this undertaking.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Crime in India. National crime records bureau. Ministry of Home Affairs, 2014 June 15. Available at: http://ncrb.nic.in.

Gandhi

Taneja

Mazumder

. Near hanging: Early intervention can save lives. Indian J Anaesth 2011; 55: 388–391.

Karanth

Nayyar

. What influences outcome of patients with suicidal hanging. J Assoc Physicians India 2005; 53: 853–856.

Penney

Stewart

Parr

. Prognostic outcome indicators following hanging injuries. Resuscitation 2002; 54: 27–29.

Nikolić

Živković

. Cervical spine injuries in suicidal hanging without a long-drop—patterns and possible underlying mechanisms of injury: an autopsy study. Forensic Sci Med Pathol 2014; 10: 193–197.

Salim

Martin

Sangthong

et al.

Near-hanging injuries: a 10-year experience. Injury 2006; 37: 435–439.

Majdan

Steyerberg

Nieboer

et al.

Glasgow Coma Scale Motor Score and pupillary reaction to predict six-month mortality in patients with traumatic brain injury: comparison of field and admission assessment. J Neurotrauma 2015; 32: 101–108.

Joseph

Alexander

Nair

et al.

Pulmonary distress following attempted suicidal hanging. Indian J Med Sci 2009; 63: 53–53.