“Marijuana has been associated with substantial adverse effects, some of which have been determined with a high level of confidence,” concluded a review by top officials at the National Institute on Drug Abuse (NIDA).

Regular and/or heavy recreational use of cannabis has been strongly linked with addictive behaviors, motor vehicle accidents, lung dysfunction, and “diminished lifetime achievement,” according to NIDA Director Nora Volkow, MD, and three of her deputies.

And such health problems as abnormal brain development, increased abuse of other drugs, and new-onset or exacerbated schizophrenia, depression, and anxiety have been identified in marijuana users, the review found with a medium level of confidence.

“As policy shifts toward legalization of marijuana, it is reasonable and probably prudent to hypothesize that its use will increase and that, by extension, so will the number of persons for whom there will be negative health consequences,” the officials wrote in the review, published in the June 5 issue of the New England Journal of Medicine.

Volkow — who has run NIDA since 2003 — and colleagues based their conclusions on findings in 77 previously published papers. They did not describe it as a comprehensive or systematic review, but rather as a survey of “the current state of science” on marijuana’s adverse health effects.

Although Volkow and colleagues specifically addressed recreational use, many of the underlying studies did not distinguish it from medical uses, and most of the review’s conclusions would likely apply to medical marijuana with high THC content when taken frequently.

Dependence and Gateway Drug Risk

Volkow and colleagues acknowledged that the addictiveness of marijuana has been “contentious,” but they insisted that the evidence falls strongly on the side of a strong propensity toward unhealthy dependence.

They mainly cited data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) and the National Survey on Drug Use and Health (NSDUH).

The NESARC analysis, from 2011, indicated that about 9% of all first-time users will become “addicted,” which Volkow and colleagues explicitly equated with having met DSM-IV criteria for dependence. The original study authors noted that the “cumulative probability of transition to dependence” was lower for marijuana than for the other three drugs they examined (nicotine, cocaine, and alcohol), but when such transition did occur, it was relatively fast.

Data from NSDUH in 2012 suggested that some 2.7 million Americans 12 and older could be considered dependent on marijuana (compared with 8.6 million meeting criteria for alcohol dependence).

The NIDA team also referred to an estimate that “about one in six” teenagers who take up cannabis become dependent on it, one of the conclusions in a 2009 Lancet review article.

Evidence that marijuana serves as a “gateway drug” — that is, a first recreational drug that leads users to initiate other, potentially more hazardous drugs such as cocaine or heroin — was less direct. Volkow and colleagues adduced rodent data indicating that cannabinoids alter reactivity in dopamine-mediated reward pathways, and that the THC component in particular “can prime the brain for enhanced responses to other drugs.”

Although the researchers asserted that “most epidemiological studies” support “increased susceptibility to drug abuse and addiction to several drugs later in life” in those starting on cannabis, their sole reference was a single twin study reported in 2004 that had numerous limitations.

Volkow and colleagues also conceded that the associations were not necessarily causal: “An alternative explanation is that people who are more susceptible to drug-taking behavior are simply more likely to start with marijuana because of its accessibility, and that their subsequent social interactions with other drug users would increase the probability that they would try other drugs,” they wrote.

Brain Development

Animal data also point to a deleterious, or at least abnormal, effect on brain development from early exposure to cannabinoids and specifically THC — but again, studies in humans have been limited to showing associations that may or may not reflect a causal relationship.

For example, Volkow and colleagues cited studies of adults who had smoked marijuana regularly during adolescence, which found “impaired neural connectivity (fewer fibers)” in certain brain regions compared with adults who were nonusers as teens.

Similar studies have found volume and activity differences in persons with a history of marijuana use versus controls.

On the other hand, these were all cross-sectional studies that did not exclude the possibility that such differences predated the marijuana use or resulted from other environmental or genetic factors that happened to correlate with cannabis use.

Mental Illness

Numerous studies have found the following mental health risks in regular marijuana users, Volkow and colleagues indicated:

• Increased likelihood of anxiety and depression
• Increased likelihood of psychosis
• Worsened symptoms of schizophrenia
• Earlier onset of psychotic events

But, they conceded, “it is inherently difficult to establish causality in these types of studies because factors other than marijuana use may be directly associated with the risk of mental illness. In addition, other factors could predispose a person to both marijuana use and mental illness.”

Motor Vehicle Accidents

Here, Volkow and colleagues did not have to rely solely on association studies. Controlled trials with driving simulators have shown clear, dose-dependent impairments associated with blood THC levels.

Pooled data from an earlier meta-analysis showed a relative risk close to 2 for accidents involving drivers who used marijuana shortly before getting behind the wheel.

Another study showed that overall motor vehicle accident risks in individuals with detectable THC were similar to those seen in persons with blood alcohol levels higher than 0.08% (although the types and severity of accidents were not necessarily the same).

Volkow and colleagues also noted that these risks of marijuana and alcohol, which are often used together, appear to be at least partly additive.

Pulmonary Health

Because marijuana is most commonly smoked, concerns about lung effects have been raised for decades. The possibilities include lung cancer, chronic bronchitis, and pulmonary infections, as well as indirect effects such as cardiovascular events — which may stem from the psychoactive compounds in the smoke or from other components.

Volkow and colleagues found suggestions but little hard evidence that smoked cannabis increases lung cancer risk. The best study so far did find an increase compared with nonsmoking controls, but adjustment for concomitant tobacco smoking erased it.

The NIDA group indicated that the evidence is stronger that cannabis smoking promotes bronchitis and increased rates of infections, mainly in the heaviest users.

Lab studies have indicated that cannabinoids may affect vascular systems, the researchers noted — CB1 receptors are found in blood vessels, for example.

Success in School and Life

Volkow and colleagues offered a plausible causal chain connecting regular marijuana use with poor school performance. For example, since the drug impairs certain cognitive functions such as short-term memory, some impact on learning would be expected.

In turn, the researchers argued, such impacts, even if relatively brief, “will interfere with the subsequent capacity to achieve increasingly challenging educational goals.”

But they cited only one publication directly addressing marijuana use and students’ grades — a review in Addiction published in 2000, in which Australian researchers Michael Lynskey, MSc, PhD, now at King’s College in London, and Wayne Hall, PhD, of the University of Queensland in Sydney, argued against a direct causal relationship. (Instead, they suggested, “early cannabis use appears to be associated with the adoption of an anti-conventional lifestyle characterized by affiliations with delinquent and substance using peers, and the precocious adoption of adult roles including early school leaving, leaving the parental home, and early parenthood.”)

The NIDA team indicated that research on possible psychosocial effects later in life has been inconsistent. Some studies, they wrote, indicated that “long-term deficits may be reversible and remain subtle.”

Others have found associations with lower income, greater reliance on public assistance, joblessness, criminality, and “lower satisfaction with life,” Volkow and colleagues wrote — though these would also be consistent with Lynskey’s and Hall’s epiphenomenon hypothesis.

Limitations

The NIDA group acknowledged that much of the human-study literature on marijuana is marred by “multiple (often hidden) confounding factors [that] detract from our ability to establish causality.”

Although those factors are often used to suggest that marijuana use is not as damaging as the crude associations might suggest, other limitations may mask even more severe risks.

For example, many of the studies were conducted more than 10 years ago, when THC concentrations in street-level marijuana were much lower than now. (Volkow and colleagues pointed to Drug Enforcement Administration data showing that THC levels have more than tripled since the 1980s.) This change in the product, the researchers suggested, puts in doubt “the current relevance of the findings in older studies … especially the studies that assessed long-term outcomes.”

They also noted the paucity of research on potential medical benefits of marijuana and its components, as well as on novel risks when prescribed for medical purposes (such as exacerbation of cognitive deficits in patients with pain states for which marijuana is sometimes recommended).