Methods Of Estimation Of Time Since Death

Article Author:
Rijen Shrestha
Article Author:
Tanuj Kanchan
Article Editor:
Kewal Krishan
Updated:
4/29/2020 12:48:03 AM
PubMed Link:
Methods Of Estimation Of Time Since Death

Introduction

Forensic pathologists are essentially required to estimate the time since death (TSD) to assist in death investigation, by limiting the period of investigation and thus helping in narrowing down the number of suspects in homicide cases. TSD also provides valuable information in establishing the timeline of events that lead to death.

Numerous researchers have examined various methods for estimating TSD in the past. While the results are encouraging and deemed useful, all the authors agree that there are extremely variable factors, often beyond the control of the examiner in real-life scenarios. In light of these findings, the authors advocate the use of multiple systems, both qualitative as well as quantitative, to ensure the reliability of the estimated TSD. 

Function

Post-mortem interval (PMI) has traditionally been classified into three stages – immediate, early and late.

Immediate PMI

In the immediate period, the body undergoes rapid biochemical and physiological changes that are primarily caused by the absence of circulation of blood and loss of regulatory mechanisms. These changes are principally detectable in the eyes and the skin. In the eyes, ‘trucking’ or segmentation of retinal blood vessels is one of the first observable signs. This sign presents as a break in the continual column of blood on ophthalmoscopic examination of the eyes, and usually occurs within half an hour and may sometimes take as long as 2 hours after death.[1] The other changes in the eyes, in the immediate post-mortem phase, include loss of intraocular pressure as well as the clouding of the cornea. The intraocular pressure decreases drastically after death and reaches 4 mmHg or less within 6 hours after death.[2] The cornea begins to cloud within 2 hours after death and usually prevents intra-ocular examination with an ophthalmoscope.[3] The skin loses its elasticity and luster within the first few hours after death and appears pale. Histological examination of the skin, however, shows no morphological changes within 6 hours PMI.[4] Other examinations show a lack of cellular or biochemical changes within 3 to 6 hours post-mortem. [5] Emptying of gastric contents is another method used for estimating the post-mortem interval. Small light meals get emptied from the stomach within 1 to 3 hours, and the time of consumption, if known – along with volume and type of meal, can be used to estimate the post-mortem interval.[6] The immediate post-mortem phase can, therefore, be termed as the post-mortem interval between somatic and cellular death, within 2 to 3 hours after death, and usually denotes a lack of discernible changes in the morphology or histo-chemistry.

Early PMI

The early post-mortem phase is probably the most important time period for the estimation of PMI as most medico-legal cases are examined in this time period. This period is also where the estimation of time since death is most relevant in establishing the timeline of events and developing a theory of circumstances of death. This period runs from 3 to 72 hours after death. The early post-mortem phase is most frequently estimated using the classical triad of post-mortem changes – rigor mortis, livor mortis, and algor mortis.

Algor Mortis

Algor mortis is the cooling of the body after death, primarily due to loss of homeostatic regulation by the hypothalamus, in conjunction with the loss of heat to the environment by conduction, convection, and radiation. Algor mortis is the most accurate method of estimating TSD in the early post-mortem phase. However, it involves a cumbersome procedure and requires intensive knowledge and research before it is accurately usable in the field; this is due to the numerous factors that affect the temperature gradient between body temperature and ambient temperature, the most inherent being the differences in the temperatures of different localities at different points of time. A rule of thumb states that there is a decrease of 1.5 degrees F every hour.[7] Several charts, formulae, and algorithms have been developed to estimate the PMI; Henssge’s nomogram being the most widely taught.[8] The estimation of TSD using algor mortis measures rectal temperatures, and while they have been consistently used, nomograms for brain-temperatures have also been developed by Brinkmann et al. in 1976 and 1978 and by Henssge et al. in 1984.[9]

Rigor Mortis

Rigor mortis is the post-mortem stiffening of muscles, caused by the depletion of adenosine triphosphate (ATP) from the muscles, which is necessary for the breakdown of actin-myosin filaments in the muscle fibers. Actin and myosin are components of the muscle fiber and form a covalent bond during contraction. The cessation of oxygen supply causes the stoppage of aerobic respiration in the cells and leads to a lack of production of ATP. Rigor mortis starts immediately after death and is usually seen in a sequence known as “march of rigor” and also called Nysten’s Law. While rigor mortis develops simultaneously in all muscle tissue in the body, voluntary and involuntary, the size of the muscle determines the perceptibility of changes by the examiner. Smaller muscles over the face – around the eyes, around the mouth, etc. are the muscles where rigor mortis first appears, followed by rigor mortis of the muscles in the hands and upper limbs and finally appears in the large muscles of the lower limbs. Rigor mortis appears approximately 2 hours after death in the muscles of the face, progresses to the limbs over the next few hours, completing between 6 to 8 hours after death.[10] Rigor mortis then stays for another 12 hours (till 24 hours after death) and then starts disappearing.[11] In the last phase of Rigor Mortis, the actin-myosin complex that has formed starts disintegrating due to proteolysis, resulting in the dissolution of the stiffness. This process begins in all the cells at the same time, however, just like with the appearance, this change is perceptible first in the smaller muscles of the face, followed by muscles of the upper limbs and finally the large muscles in the lower limbs. Rigor mortis generally disappears in 36 hours after death, followed by a phase known as secondary flaccidity.

Livor Mortis

The final change in the classical triad is livor mortis, which is the purplish-blue discoloration of the skin in the dependent parts of the body, due to collection of blood in skin vessels, caused by gravitational pull. Hypostasis develops as spots of discoloration within half an hour to 2 hours, these spots then coalesce to form larger patches, which further combine to form a uniform discoloration of the dependant parts of the body that has not been subject to pressure, which appears from 6 to 12 hours. The discoloration becomes ‘fixed’ after a certain period, owing to the disintegration of blood cells and the seepage of hemoglobin. This fixation is confirmed by applying pressure with thumbs and is traditionally used to denote a PMI greater than 12 hours.[12] This method of estimation of PMI required an objective and modern approach, leading to the development of colorimetric methods for the estimation of PMI from livor mortis.[13]

Histo-morphological analysis

Other methods of estimating TSD in the early phase include histo-morphological and Bio-chemical analysis. Total and differential blood counts, as well as the microscopic morphological examination of blood, has been described as a method for estimation of the TSD. All blood cells were not identifiable beyond 84 hours after death. Similarly, blood cell counts were also found to decrease beyond 84 hours after death.[5] Histological studies of the skin have shown that degenerative changes appear in the skin 6 hours after death and first appears as vacuolation of the corpus basale and spinosum. Dermo-epidermal separation is seen 9 hours after death, while dermis showed rarefaction and disintegration 6 and 18 hours after death, respectively. The glycogen in the basal membrane of the sweat glands, the secretory cells cytoplasm as well as duct cells gets depleted within 3 hours PMI and leads to PAS-negative cells on histology. The basal membrane, however, continues to show a magenta staining up to 18 hours post-mortem. The eccrine sweat glands show vacuolation after 3 to 4 hours of PMI, and cells appear to have completely disintegrated 15 hours after death. The sebaceous glands appear normal till 18 hours post-mortem, seen as a separation of the layers and disintegration of hair papilla.[4] Studies have also shown that pleocytosis can be used to estimate the PMI using a polynomial equation of the third order. The cells are primarily lymphocytes with a significant fraction of macrophages, which become vacuolated and unidentifiable after 12 hours.[14]

Biochemical assessment

Biochemical blood assessment is non-significant in the immediate post-mortem phase due to the lack of cellular death. On the other hand, cellular death makes biochemical blood assessment in the early phase extremely difficult. Also, there is the redistribution of electrolytes from the cells into the plasma and serum, resulting in varying changes in the levels of these electrolytes. These variations and their implications are studied in the emerging field of thanato-chemistry. The biochemical assessment has been useful for estimating PMI from vitreous humor, synovial fluid, pericardial fluid, urine, and cerebrospinal fluid. Numerous factors, however, need to be accounted for when examining the PMI based on biochemistry including, but not limited to, age, gender, biological background, lifestyle, cause of death, and a whole range of other intrinsic and extrinsic factors. Only a few biochemical markers (out of 388) were found to have had sufficient investigation with these considerations – namely potassium, sodium, urea as well as chloride, magnesium, hypoxanthine, and cardiac troponin T.  Assessment for their potential for use was found to be alarming, with 0 (zero) biochemical markers being judged to have had suitable research and suitable for use. Six were found to be suitably researched but not suitable for practical use. Meanwhile, 18 were found to have been poorly investigated and not suitable for application, and a further 364 biochemical markers did not have sufficient information.[15]

Supra-vital reactions

Supra-vital reactions have also been proposed as a means of estimation of PMI. The determination of the supra-vitality period, therefore, can help assist in the estimation of PMI. For this method, Madea defines the PMI into four stages - the latency period, where despite stoppage of circulation, the tissue still performs aerobic respiration till the depletion of its stores – the survival period, where there is loss of tissue function, but they can be re-activated using external stimuli, e.g., electrical stimulation of nerves – the resuscitation period, where the ability of the tissue to recover is completely lost, –  and the supra-vital period. Madea defines supra-vitality as the survival period of tissue after complete, irreversible ischemia. This concept states that the survival period encompasses the latency period, the resuscitation period encompasses both the latency period and the survival period and supra-vitality period includes all the other three. Supra-vitality is also different from the resuscitation period in that the tissue is excitable irrespective of recovery of function. As an example, the resuscitation period of skeletal muscle is approximated to be 2 to 3 hours, but the supravital period in some cases may extend to 20 hours. Similarly, cardiac muscles have a resuscitation period of 3.5 to 4 min, while the supravital period may extend up to 2 hours.[16] A method for estimating the PMI was developed using the electric excitability of Orbicularis oculi using surface electrodes. A ratio of relaxation time and maximum force, called force-related relaxation time, was found to be reliable for estimating the PMI. It is also important to consider the super-sensitivity of tissue in the immediate post-mortem phase, called Zsako’s phenomenon. The supra-vital reaction, therefore, examines the idio-muscular or local contraction and not the contraction of the entire muscle.[9]

Late PMI

The late post-mortem phase is the period when the body tissue starts disintegrating and is primarily describable as decomposition or putrefaction, adipocere formation, mummification, or skeletonization. Complex tissue in the body starts disintegrating into simpler molecular forms as a result of the breakdown of the tissue by the body’s enzyme or bacteria as well as bacteria that colonize the remains after death. The body primarily undergoes decomposition or putrefaction, resulting in greenish discoloration, bloating due to gas formation, and liquefactive necrosis. The decomposition of remains is dependent on the climate, the season, body weight, and clothing. Decomposition can divide into five stages – fresh, early decomposition, advanced decomposition, skeletonization, and extreme decomposition.

Fresh phase

The fresh phase can start as early as 24 hours and as late as 7 days after death, especially in colder winter months. This phase shows no insect activity other than the deposition of blow fly-eggs in the cavities and areas of tissue dehiscence. Egg deposition has been documented in living patients, especially in immobile and debilitated subjects. [17]

Early decomposition phase

The early decomposition phase begins with the onset of skin slippage and hair loss. These changes usually begin from the first day after death to up to five days post-mortem. Maggots also begin to appear on the body, starting from the second-day post-mortem; the body appears grayish-green and marbling present (some parts of the body may still appear pinkish). The right iliac fossa is the first body part to show greenish discoloration and may be seen as early as the second-day post-mortem. This is due to the relatively superficial position of the caecum. The extremities appear brownish with the drying of the skin, especially over the fingers, nose, and ears, usually beginning on the second post-mortem day; the body appears greenish with distinct bloating. The greenish discoloration, that started at the right iliac fossa progresses to encompass the entire abdomen, with concurrent bloating of the abdomen, which again may start on the second day. The bloating advances to the rest of the torso and subsequently the body, resulting in crepitations over the entire body on handling. This phase is also associated with purging – the release of decomposition fluid from the orifices – and a strong disagreeable odor. Bloating may be seen as early as three days after death and usually subsides by the second-week post-mortem, due to disruption of the abdominal gases; The body appears blackish-green by the second-week; and finally, the body appears brownish-black with the leathery appearance of skin. This stage is usually seen until the end of the first month but may be prolonged to as long as two months. The underlying tissue also appears darkened with the texture changing to a viscous, slimy paste. Between the tenth day and the end of the first month, maggot activity continues under the leathery skin, with the skin further desiccating to form a hardened leathery shell, with loss of underlying soft tissue.

Advanced decomposition phase

The advanced decomposition phase begins with the appearance of loose sagging skin and the collapse of the abdominal cavity. The body also shows extensive maggot infestation. These changes usually appear at the fourth-day post-mortem but may begin as late as ten days after death. Loss of soft tissue, including the loss of the desiccated leathery skin, results in exposure of less than half of the skeletal material. This phase is usually associated with the presence of pupal cases, as well as the appearance of molds over the body and clothing; this usually occurs in the second month but may occur six to nine months post-mortem. Desiccation of the outer skin could accompany the structural retention of internal organs, or be accompanied by autolysis and loss of internal organs. Decomposition may progress rapidly in buried remains or in remains left in an environment with high humidity, resulting in extreme maggot activity, accelerated autolysis, and could progress directly to skeletonization or adipocere formation, without desiccation and mummification of the skin and outer tissue.[18] The remains may undergo either saponification or desiccation, called adipocere formation and mummification, respectively, depending on the environment and conditions present. The presence of a warm, humid environment that lacks oxygen favors the development of adipocere – a waxy substance that results from anaerobic bacterial hydrolysis of body fat. The primary organism responsible for adipocere formation is Clostridium perfringens, causing causes aggregation of crystal of fatty acid, resulting in loss of cellular detail as well as the loss of epidermis.[19] The formation of adipocere and the time duration depends primarily on the pH, temperature, moisture, and lack of oxygen in the environment.[20]

Skeletonization phase

The skeletonization phase results in exposure of more than half of the skeletal elements, that could still demonstrate soft tissue that is still attached. However, skeletonization is usually associated with desiccated tissue or adipocere over less than half of the body. The desiccated tissue most commonly appears at muscular or ligamental attachments along the vertebral column as well as ends of long bones. Meanwhile, adipocere is commonly seen over the thighs, having high subcutaneous fat deposits. This stage appears two months after death, although it usually presents between two and nine months post-mortem. Continuation of decomposition leads to exposure of all osseous material, with only some greasy material left behind or exposing dry bones; this is usually seen after six months of exposure, although it has been reported to have occurred as early as the third week. This stage can last for years if the elements are not exposed, as is seen in buried remains or remains found indoors.

Extreme decomposition phase

The phase of extreme decomposition is seen only in remains that have been exposed to the environment and leads to erosion of the skeletal elements. This erosion begins with the process of bleaching of bones and is commonly seen six months after exposure, although it has been documented to appear as early as two months after death and as late as two and a half years post-mortem. The skeletal elements undergo further degeneration of the cortical structure, resulting in a metaphyseal loss in long bones and exposure of the cancellous part of spongy bones; this is seen commonly between a year to a year and a half after death, although it has been reported to have occurred as early as the fourth month. The metaphyseal loss was reported to have occurred at PMI of five and a half years.[18]

Forensic Entomology

Forensic entomology analysis has been a routine practice for the estimation of PMI in the early and late post-mortem periods. There are two methods of estimation using forensic entomology – based on succession and based on development. In a succession-based approach, A suitable succession-model is chosen for use, one that corresponds to the environmental conditions, including the circumstances of death. Therefore, forensic research into the effect of environmental factors on decomposition and insect succession is needed.[21] Mañas-Jordá demonstrated that different taxa were found to be prevalent based on environmental conditions. The species diversity, as well as the number of individuals, were examined. They detected no species association with Stage I and II of decomposition, three species [Compsomyiops spp (Diptera: Calliphoridae) and C. irazuana, and Megagrapha sp1 (Hybotidae)] associated with Stage III, two species [Mesosphaerocera sp1. and Fannia sp1. (Fanniidae)] associated Stage IV and one species [Stilpon sp1 (Hybotidae)] associated with Stage V at Huitepec Natural Reserve. At the City of San Cristóbal de las Casas, four species [Prochyliza brevicornis (Melander; Diptera: Piophilidae), C. latifrons, Lucilia mexicana (Calliphoridae) and Compsomyiops spp (Macquart)] were found to be associated with Stage II, three species [Synthesiomyia nudiseta (Vanderwulp), Musca sp1, and Hydrotaea sp1(Diptera)] were found to be associated with Stage III, only one species Chrysomya rufifacies (Macquart; Diptera: Calliphoridae) and Fannia sp1 associated with Stage IV and lastly, Stilpon sp1 was found to be associated with Stage V.[22]

The development-based approach looks at the presence of different stages of the insect on the body as well as in the surrounding area, to help estimate PMI. Matuszewski used L. caesar (Diptera: Calliphoridae), Thanatophilus sinuatus, and N. littoralis (Coleoptera: Silphidae) in his research to demonstrated that the presence of a developmental stage and absence of the subsequent developmental stages of carrion insects, could be used in conjunction with the estimation of their pre-appearance interval (PAI) to develop a reliable estimation of PMI. It is, therefore, essential to establish known PAI values for different insects in the environment under examination.[23]

Molecular assessment

Recent advances in molecular biology have led to various advances in the estimation of PMI. The degeneration of mRNA, DNA, and proteins are evaluated and can be used to estimate the PMI. RNA transcripts were found to be the most relevant due to their rapid degeneration and temporal correlation. Multiple studies demonstrated a linear correlation between PMI and degeneration. This correlation was found to be temperature and tissue-dependent.[24]

A study from Porto, Portugal examined 11 gene transcripts for correlation with TSD. 8 murine tissues were divided into three groups based on the stability of the RNA - first group (I) comprising of tissue samples from the heart, spleen and lung, second group (II) consisted of femoral quadriceps, liver and stomach and the third group (III) Pancreas and skin. Samples from the groups I and II were serially analyzed. The analysis showed RNA degeneration was time-dependent for the entire 11 hours, although no statistical significance was demonstrable for the first four hours. Researchers selected 11 genes for quantitative PCR analysis. While RNA in the heart was found to be most stable, it showed no correlation with PMI. A total of six genes were found to correlate with PMI, four in the femoral quadriceps (Actb, Gapdh, Ppia, and Srp72) and two genes in the liver (Alb and Cyp2E1). Mathematical models were developed to estimate PMI with an error mean of 51.4 minutes.[25]

Issues of Concern

All studies have shown that numerous factors were beyond the scope and control of the research, which invariably affects the rate of development of post-mortem changes in real-life scenarios; this requires the use of multiple methods of estimating PMI to help establish a reliable timeline.

This is further complicated by slight variations in sample collection and processing, combined with the individual variations in the temperatures at time of death (body and ambient), clothing worn by the deceased, circumstances of death (including cause of death, preservation after death, exposure to the elements etc.), methods of retrieval as well as storage and preservation of samples before analysis all affect the estimation of PMI in a deceased.

Clinical Significance

Information about time since death is an essential feature of postmortem reports issued by the forensic pathologists. The estimation of post-mortem interval provides useful information regarding the time of death, and hence, the same is argued upon by the defense in the court of law. Clinical expertise is warranted to ensure that the postmortem changes are well- interpreted and inferences get drawn correctly. The estimated time since death or the PMI may range in hours and sometimes in days when much time has elapsed after death, which may not always be helpful to law enforcement, who would require a narrower range of estimates to help in the investigation. The need for further research is, thus, emphasized.

Enhancing Healthcare Team Outcomes

Estimating the time since death is among the pertinent challenges faced by forensic pathologists and law enforcement agencies all over the world. While this is primarily the bailiwick of the forensic pathologist or medical examiner in collaboration with law enforcement, as is demonstrated in the preceding discussion, this requires collaborative efforts from an interprofessional team, including bio-chemist, pathologists, anthropologists, in addition to the already mentioned forensic pathologist and law enforcement personnel. Other health care practitioners such as clinicians and nursing need to have a basic understanding of the processes, so they can provide input when requested, and be aware not to damage any evidence accidentally.


References

[1] Jaafar S,Nokes LD, Examination of the eye as a means to determine the early postmortem period: a review of the literature. Forensic science international. 1994 Feb;     [PubMed PMID: 8175090]
[2] Van den Oever R, A review of the literature as to the present possibilities and limitations in estimating the time of death. Medicine, science, and the law. 1976 Oct;     [PubMed PMID: 979556]
[3] Wróblewski B,Ellis M, Eye changes after death. The British journal of surgery. 1970 Jan;     [PubMed PMID: 4904161]
[4] Bardale RV,Tumram NK,Dixit PG,Deshmukh AY, Evaluation of histologic changes of the skin in postmortem period. The American journal of forensic medicine and pathology. 2012 Dec;     [PubMed PMID: 21897193]
[5] Babapulle CJ,Jayasundera NP, Cellular changes and time since death. Medicine, science, and the law. 1993 Jul;     [PubMed PMID: 8366783]
[6] Madea B, Methods for determining time of death. Forensic science, medicine, and pathology. 2016 Dec;     [PubMed PMID: 27259559]
[7] Nokes LD,Flint T,Williams JH,Knight BH, The application of eight reported temperature-based algorithms to calculate the postmortem interval. Forensic science international. 1992 May;     [PubMed PMID: 1639277]
[8] Henssge C, Death time estimation in case work. I. The rectal temperature time of death nomogram. Forensic science international. 1988 Sep;     [PubMed PMID: 3192144]
[9] Henssge C,Madea B, Estimation of the time since death in the early post-mortem period. Forensic science international. 2004 Sep 10;     [PubMed PMID: 15364387]
[10] Anders S,Kunz M,Gehl A,Sehner S,Raupach T,Beck-Bornholdt HP, Estimation of the time since death--reconsidering the re-establishment of rigor mortis. International journal of legal medicine. 2013 Jan;     [PubMed PMID: 22015934]
[11] BATE-SMITH EC,BENDALL JR, Factors determining the time course of rigor mortis. The Journal of physiology. 1949 Dec 15;     [PubMed PMID: 15406381]
[12] Kaatsch HJ,Schmidtke E,Nietsch W, Photometric measurement of pressure-induced blanching of livor mortis as an aid to estimating time of death. Application of a new system for quantifying pressure-induced blanching in lividity. International journal of legal medicine. 1994;     [PubMed PMID: 8038114]
[13] Vanezis P, Assessing hypostasis by colorimetry. Forensic science international. 1991 Dec;     [PubMed PMID: 1783332]
[14] Wyler D,Marty W,Bär W, Correlation between the post-mortem cell content of cerebrospinal fluid and time of death. International journal of legal medicine. 1994;     [PubMed PMID: 8038112]
[15] Meurs J,Krap T,Duijst W, Evaluation of postmortem biochemical markers: Completeness of data and assessment of implication in the field. Science     [PubMed PMID: 30798866]
[16] Madea B, Importance of supravitality in forensic medicine. Forensic science international. 1994 Dec 16;     [PubMed PMID: 7860008]
[17] Keh B, Scope and applications of forensic entomology. Annual review of entomology. 1985;     [PubMed PMID: 3882048]
[18] Galloway A,Birkby WH,Jones AM,Henry TE,Parks BO, Decay rates of human remains in an arid environment. Journal of forensic sciences. 1989 May;     [PubMed PMID: 2738563]
[19] Byard RW, Adipocere-The Fat of Graveyards. The American journal of forensic medicine and pathology. 2016 Sep;     [PubMed PMID: 27356012]
[20] Jopp-van Well E,Augustin C,Busse B,Fuhrmann A,Hahn M,Tsokos M,Verhoff M,Schulz F, The assessment of adipocere to estimate the post-mortem interval - a skeleton from the tidelands. Anthropologischer Anzeiger; Bericht uber die biologisch-anthropologische Literatur. 2016;     [PubMed PMID: 27189778]
[21] Jarmusz M,Bajerlein D, Decomposition of hanging pig carcasses in a forest habitat of Poland. Forensic science international. 2019 Jul;     [PubMed PMID: 31075565]
[22] Mañas-Jordá S,León-Cortés JL,García-García MD,Caballero U,Infante F, Dipteran Diversity and Ecological Succession on Dead Pigs in Contrasting Mountain Habitats of Chiapas, Mexico. Journal of medical entomology. 2018 Jan 10;     [PubMed PMID: 29186358]
[23] Matuszewski S, A general approach for postmortem interval based on uniformly distributed and interconnected qualitative indicators. International journal of legal medicine. 2017 May;     [PubMed PMID: 28054103]
[24] Scrivano S,Sanavio M,Tozzo P,Caenazzo L, Analysis of RNA in the estimation of post-mortem interval: a review of current evidence. International journal of legal medicine. 2019 Jul 18;     [PubMed PMID: 31317317]
[25] Sampaio-Silva F,Magalhães T,Carvalho F,Dinis-Oliveira RJ,Silvestre R, Profiling of RNA degradation for estimation of post mortem [corrected] interval. PloS one. 2013;     [PubMed PMID: 23437149]