A person who reports having used methamphetamine and sildenafil at some point in life has not necessarily taken them together. That distinction matters when reading Fisher and colleagues’ study of men recruited through HIV prevention and testing programs in Long Beach, California. The study identified associations worth investigating, but it did not experimentally administer a drug combination or establish a pharmacological interaction.
What population did the researchers actually study?
The investigators collected information from 1,839 men between May 2001 and July 2007. Participants were recruited through three prevention or testing programs, rather than sampled randomly from all residents. Their ages ranged from 15 to 88 years. The article appeared online in 2009 and in the 2011 volume of Archives of Sexual Behavior. These are historical observations, not estimates of present-day prevalence.
Missing information reduced the multinomial analysis to 1,794 complete cases. Within that analytic sample, 762 reported neither drug, 715 methamphetamine only, 118 sildenafil only, and 199 both. The last category represented 11.1% of complete cases. It should not be presented as an 11.1% prevalence of simultaneous methamphetamine–sildenafil ingestion in the general population.
Why does the exposure definition matter?
The interview instruments asked about lifetime drug use, as well as more recent frequency and sexual behavior. The four-group classification was based on reported use histories. Membership in the “both” category alone cannot establish the date, dose, route, sequence, or temporal overlap of the two drugs. It also cannot establish the identity or concentration of a drug in a specimen from a particular encounter.
The paper separately examined reported methamphetamine use immediately before or during sex. It reported 24% in the methamphetamine-only group and 33% in the group reporting both drugs, with an analysis denominator of 542 and p = .0502. That result sits just above the conventional .05 threshold; describing it as an unequivocally significant difference would erase a relevant qualification. It also does not transform every lifetime report into documented same-event co-use.
The infection comparisons had different denominators
Table 1 reported laboratory results for subsets of the sample. The hepatitis B comparison included 631 tested participants and the HIV comparison 1,021. The two syphilis-test comparisons included 580 and 582. Those denominators differ from both the 1,839 recruited participants and the 1,794 complete cases used in the regression.
For example, the table reported HIV positivity of 9.5% in the group reporting both drugs and 3.8% in the neither-drug group. These are proportions among tested members of those groups. They are not prospective probabilities of infection after a dose, and the table does not justify dividing all infection counts by the full recruitment sample.
What did the statistical model predict?
The multinomial model compared drug-use categories with the neither-drug reference category. In the both-drugs comparison, lifetime GHB/GHL use had an odds ratio of 10.8, with a 95% confidence interval of 4.30–28.80. The direction of that model is important: the outcome was membership in a reported drug-use group. This number is not a 10.8-fold increase in HIV risk caused by GHB, nor a measurement of the toxicity of a three-drug combination.
The overlapping histories of cocaine, ecstasy and other drug use also make it difficult to attribute an observed behavioral or health difference to methamphetamine alone.
What can this evidence establish for an individual?
It supports an association between particular reported exposure histories and behavioral or infection markers in a selected service population. Self-report, recruitment through prevention programs, incomplete testing, and the absence of established temporal sequence constrain causal interpretation. Infection could precede a reported drug-use pattern; shared social or behavioral factors could contribute to both.
For a forensic evaluation, the study supplies background for asking precise questions about exposure and chronology. It does not establish an individual’s intoxication, sexual conduct, consent, source of infection, or impairment. Those conclusions require evidence about the actual person and event.
Source
Fisher DG, Reynolds GL, Ware MR, Napper LE. Methamphetamine and Viagra Use: Relationship to Sexual Risk Behaviors. Archives of Sexual Behavior. 2011;40:273–279. Published online March 28, 2009. doi:10.1007/s10508-009-9495-5. Historical observational survey; no controlled drug administration.
Leave a Reply