I came across this interesting feature written by Stanton Peele.  I've selected some important paragraphs.  For full text and related info, visit http://www.peele.net/lib/heroinoverdose.html

© Copyright 1996-2009 Stanton Peele. All rights reserved.

People rarely die from heroin overdoses — meaning pure concentrations of the drug which simply overwhelm the body's responses. People do die while consuming heroin — but the overdose myth may actually make such deaths more, rather than less, likely.

The first popular source to tell us about the myth of heroin overdose was the classic 1972 Consumer Union Report, Licit & Illicit Drugs, written by Edward M. Brecher. Brecher pointed out that, when street doses of heroin were far purer than they are today (China Cat and black-tar heroin scares notwithstanding), drug overdoses were practically unknown.

Brecher noted that heroin overdoses began to be reported in New York City after World War II, and accelerated into the 1970s. Yet the average purity of a street dosage prior to the War was 40 times the concentration of a 1960s dose.

Research at the Jefferson Medical Center in Philadelphia in the 1920s showed that addicts could tolerate up to a ninefold increase in the concentration of their standard, already large, dose. These researchers estimated that a toxic dose of heroin would be at least 500 milligrams for nonusers and 1800 milligrams for addicts.

In the 1960s, New York City Medical Examiners Drs. Milton Helpern and Michael Baden studied heroin addict deaths. Heroin found near dead addicts was not unusually pure and their body tissues did not show especially high concentrations of the drug. Although the addicts typically shot up in groups, only one addict at a time died. Furthermore, the dead addicts were experienced rather than novice users and therefore should have built up tolerance to large doses of heroin.

The best guess as to what was killing these addicts (aside from general infection, illness, and malnutrition) were the impurities in the drug, such as quinine, which produced adverse reactions in some injectors. A related likelihood which is more evident today is the mixture of drugs, or of drugs and alcohol.

Street lore among heroin addicts typically eschewed drinking alcohol with heroin as a potentially deadly combination. Today, drug cocktails as well as drinking while shooting up are common. The majority of drug deaths in an Australian study, conducted by the National Alcohol and Drug Research Centre, involved heroin in combination with either alcohol (40 percent) or tranquilizers (30 percent).

If it is not pure drugs that kill, but impure drugs and the mixture of drugs, then the myth of the heroin overdose can be dangerous. If users had a guaranteed pure supply of heroin which they relied on, there would be little more likelihood of toxic doses than occur with narcotics administered in a hospital.

But when people take whatever they can off the street, they have no way of knowing how the drug is adulterated. And when they decide to augment heroin's effects, possibly because they do not want to take too much heroin, they may place themselves in the greatest danger.

. Death by asphyxiation in one's vomit is common among people who mix alcohol with drugs, which often occurs at parties. Alcohol is more likely to cause people to puke, while additional drugs make the intoxicated individuals less able to stir themselves awake.

According to Dr. Laurence Gruer, addictions coordinator for the Greater Glasgow Health Board, "It is actually rare to find someone has died taking heroin alone — it has almost inevitably been taken as part of a cocktail with [tranquilizers] Temazepam or Valium."

Public officials can generally say any bad thing they want about illegal drugs. And they feel no doubt that labeling deaths as overdoses should scare most young people away from drugs. But this message may not have the desired effect. And its unintended consequences can be deadly. As the Australian National Research Centre made clear: "Both heroin users and service providers need to be disabused of the myth that heroin overdoses are solely, or even mainly, attributable to fluctuations in heroin purity."