Post-mortem examinations of all occupants of the aircraft were conducted by the United States Armed Forces Institute of Pathology (AFIP) under the control and supervision of CASB investigators. In addition, toxicology testing for the presence of carbon monoxide (CO), hydrogen cyanide (HCN) and common drugs was conducted on tissue and fluid specimens obtained during autopsy. Blood samples were obtained from thoracic vessels or the heart where possible. When these vessels were disrupted and blood unavailable elsewhere, specimens were obtained from pooled thoracic blood. The toxicology testing was conducted at the Civil Aviation Medical Unit (CAMU) of the Department of National Health and Welfare located in Toronto, Ontario.

Prior to autopsy, all accident victims were radiographed and examined carefully for injuries indicative of explosive blast effects and/or fragmentation associated with the detonation of an explosive device. The radiographs and bodies were specifically examined for trace evidence such as scrapnel and/or identifiable portions of an explosive device. No characteristic injury patterns, trace evidence, or portions of an explosive device were detected.

All three flight crew members sustained multiple fatal injuries on impact. No evidence of causal or contributory pre-existent disease or other physical problems that would affect the flight crew's judgement or performance was detected through autopsy of the three night crew members.

Toxicological test results were negative for the presence of CO in the case of the captain and flight engineer. Toxicological test results for the presence of HCN were negative in the case of the captain and, in the case of the flight engineer, revealed an HCN level in the blood of 0.01 milligrams per 100 millilitres (mg%). Both the captain and flight engineer tested positive for caffeine and salicylic acid (ASA/Aspirin). In the case of the first officer, insufficient fluid samples were available to test for the presence of CO and HCN. Test results for the presence of common drugs were negative.

Complete autopsies were performed on three of the five flight attendants. The remaining two flight attendants were not autopsied in deference to family requests made on religious grounds. However, both remains were radiographed and external observations made.

The five flight attendants sustained multiple traumatic injuries. Specimens for toxicological testing were obtained from the remains of three night attendants. Toxicological tests for the presence of CO were positive in the case of two flight attendants. Measured levels of CO were 21 per cent and 5 per cent saturation. Toxicological tests for the presence of HCN revealed a 0.12 ma% HCN level in the blood of the night attendant with the 21 per cent CO level, the remaining tests for HCN were negative. Tests for common drugs were positive for caffeine in all three cases and positive for salicylic acid (ASA/Aspirin) in two cases.

All 248 passengers sustained fatal injuries as a result of impact and/or the result of fire. Toxicological tests determined positive values of CO in 69 of the 189 passenger samples available for testing. Toxicological tests determined positive values of HCN in 158 of the 187 passengers where measures of HCN were available.

All but two of the blood samples in which a positive CO value was detected also gave evidence of positive HCN findings. However, there was no correlation between level of CO and level of HCN.

An extensive analysis of lung tissue was undertaken to identify possible evidence of explosive blast effects and/or evidence of inhalation of hot air, toxic gases and/or soot. The results of this analysis were inconclusive. The effects of an explosive blast wave were considered indistinguishable from the effects of trauma due to decelerative forces, flying debris, and structural collapse of the aircraft. Similarly, it was not possible to distinguish between the pulmonary effects of a pre-impact or post-impact fire.

To use respiration of products of combustion as indicators of the timing of a fire, it is necessary to assess the likelihood of a victim surviving the impact. Where a significant number of victims unlikely to have survived an impact show evidence of respiration of the products of combustion, this would indicate there was a fire before impact. Where a significant number of victims likely to have survived an impact show evidence of respiration of combustion products and very few of those unlikely to have survived the impact exhibit traces of the respiration of combustion products, this would be a strong indication of a post-impact fire only.

A complete review of pathological examination results was undertaken for the CASB by forensic pathologists from the University of British Columbia and University of Toronto, and AFIP representatives. The primary purpose of this review was to estimate the time interval from injury to death for each victim. Injuries were coded according to severity using a modification of the approach taken by the Abbreviated Injury Scale, a commonly used injury severity index developed by the American Association of Automotive Medicine. Injury pattern coding was completed without reference to CO or HCN levels. The time intervals from injury to death were estimated as follows:

In six cases, it was not possible to estimate the time interval from injury to death.

The high numbers of victims with positive levels of HCN detected in toxicological examinations caused CASB investigators to conduct an examination of the HCN phenomenon. An extensive literature review revealed that conflicting opinions exist among forensic experts with respect to the mechanisms whereby measurable levels of HCN can enter the blood of an accident victim. Bacteria action, physical decay, freeze-thaw cycles, and direct contamination have all been cited as factors that can cause elevation of HCN levels in blood samples. As a result of this review, CASB investigators conclude that four primary mechanisms leading to measurable levels of HCN must be considered when assessing levels of HCN in accident victims.

1. Background

   HCN may be found in low concentrations in the blood of normal people. Smokers can have levels as much as twice the "normal" level.

   Neo-formation

   A number of processes can produce HCN in the body. These processes are very complex and their impact is not well understood. They include bacterial activity, the breakdown of thyocyanate, and the production of HCN as a result of the burning and subsequent freezing of the body. It appears that, under certain conditions, these processes can result in wide variations in HCN levels.

   Contamination

   In victims who suffer penetrating chest wounds, HCN can be directly introduced into blood in the chest cavity through direct contact with combustion products. HCN does not combine with haemoglobin to form a virtually impermeable barrier at the blood/air interface as does CO. Rather, it continues to diffuse into the pooled blood until a point of equilibrium is reached.

   Respiration of Combustion Products

   HCN is given off as a product of combustion of many of the materials found in aircraft interiors. Victims who breath in air contaminated with HCN will show elevated levels of HCN in the blood. Depending on the concentration of HCN in the air and the amount of con- taminate air breathed in, levels of HCN in the blood can be quite high.

The role that factors other than respiration of products of combustion play in elevating HCN levels in accident victims' blood is illustrated by new procedures adopted by CAMU. Prior to this accident, the threshold level used by CAMU when reporting the presence of HCN was 0.01 mg%. However, since the accident, CAMU has revised this threshold upward to 0.02 ma% in response to the common detection of HCN in accident victim blood samples where no exposure to fire occurred. Application of this new threshold value in blood sample analyses from this ac- cident would result in 30 fewer cases of positively reported HCN levels.

A statistical analysis was performed to identify and correlate the mechanisms involved in produc tion of the HCN levels observed in the victims of this accident and to correlate the evidence regarding the possibility of a pre-impact fire as illustrated by CO and HCN levels, and soot traces found below the trachea in micropathological examinations. The statistical analysis indicated that more than one mechanism was involved in the production of the observed HCN levels. The most important mechanism was determined to be inhalation of products of combustion. Contamination of the blood through direct entry via chest wounds and the production of HCN through exposure to tobacco were also determined to have significant effects, both singly, and in combination.

In the 27 cases where survival for any time was considered to be unlikely and blood samples were available for toxicological testing, 24 cases showed a zero level of CO. The three cases which showed non-zero levels of CO were all below 25 per cent saturation. Where survival was estimated to be less than 30 seconds, 36 of 41 cases where blood samples were available showed zero levels of CO. In the 125 cases where survival was estimated at between 30 seconds and five minutes and blood samples were available, CO was found in 62 cases.

In the 39 cases where soot was found below the trachea, 38 were among victims where the estimated survival time was between 30 seconds and 5 minutes. In the remaining case, survival time was estimated at less than 30 seconds. There were no cases of soot found below the trachea where survival time wac ectimated to be zero.

In the 160 cases where measurable HCN was detected, 51 cases involved an estimated survival of less than 30 seconds and/or zero. However, 45 of these cases showed evidence of severe chest wounds and/or exposure to tobacco products. The remaining six cases were victims for which the survival time was estimated to be less than 30 seconds but not zero. There were no cases of positive HCN levels where survival time was estimated to be zero, and the confounding factors of chest wound and/or exposure to tobacco products were not also present.

The statistical analysis concluded that the comparison of survival time estimates with evidence of respiration of combustion products strongly supported the proposition that a number of accident victims survived the initial impact and died in the post-impact fire and that the comparison did not suppon a pre-impact fire scenario.

Mechanism of death was determined in 247 of the 256 cases. In the remaining nine cases, post-mortem disruption prevented determination of the mechanism of death. Mechanisms of death were determined as follows:

1. One hundred and seventy-five victims died as a direct result of injuries sustained in the impact.

2. Thirty-one victims died as a direct result of the inhalation of products of combustion. Impact injuries played no material role in their deaths.

3. Forty-one victims died as a result of the combined effects of the inhalation of products of combustion together with the injuries sustained in the impact.