Prescription stimulant use in the United States increased over the past decade,1 approximately doubling from 2002 to 2005.2 A potential adverse consequence of prescription stimulant use is small but significant increases in systolic (2–4 mm Hg) and diastolic (1–3 mm Hg) blood pressure (BP) and heart rate (HR) (4–6 beats per minute [bpm]).3–11 Given the relationship between BP and risk of cardiovascular events,12 concern exists that prescription stimulant use among adults may increase the incidence of adverse cardiovascular outcomes.13–16 Current US Food and Drug Administration labeling for prescription stimulants cite “moderate to severe hypertension” as a contraindication,17 warn about use when increases in BP or HR could be medically compromising,17,18 and caution about use “even with mild hypertension.”19 However, no guidance is provided for prescribing these medications to patients who are nonhypertensive.
Previous research has focused on the magnitude of hemodynamic changes associated with prescription stimulants and not on predictive patient risk factors. The present study’s aim was to determine what baseline variables predict subsequent increases in BP. Data from the National Drug Abuse Treatment Clinical Trials Network multisite smoking cessation trial of nicotine-dependent adults with attention deficit hyperactivity disorder (ADHD) using transdermal nicotine were analyzed. The parent study hypothesized that participants treated adjunctively with osmotic-release oral system methylphenidate (OROS-MPH), relative to placebo, would be more successful in smoking cessation. However, no significant difference in smoking cessation between treatment groups was observed.20 Because adults with ADHD (42%) smoke more than the general population (26%),21 and hemodynamic changes were constantly assessed during OROS-MPH treatment, this trial presented an ideal opportunity to study risk factors associated with OROS-MPH–mediated increases in BP.
Trial Participants A full description of the trial has been reported previously.20 In brief, 255 adults (aged 18–55 years) were recruited. Inclusion criteria included current smoking status and a diagnosis of ADHD.22,23 Persons with BP >135/85 mm Hg or a HR >90 bpm on 3 consecutive pre-randomization visits were excluded. Later, exclusion criteria were changed to requiring only two consecutive pre-randomization visits with elevated measures, and BP >130/80 mm Hg or HR >88 bpm for participants 40 to 55 years old. Other exclusion criteria included current Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnosis of substance abuse/dependence, positive urine toxicology screen, ADHD treatment with stimulants in the past 30 days, and medical conditions deemed unsafe (eg uncontrolled hypertension). After randomization, trial participants developing BP >140/90 mm Hg on two consecutive clinic visits, or initiating an antihypertensive drug, could have their OROS-MPH dosage decreased. Notably, no trial participants dropped out because of adverse events.
Study Participants For this secondary analysis, 230 participants (of the original 255) were analyzed. Seven trial participants with elevated baseline systolic BP (SBP) (≥140 mm Hg), diastolic BP (DBP) (≥90 mm Hg), or HR (≥100 bpm) were excluded. Also excluded were 14 participants taking an antihypertensive drug prior to randomization. Four participants were unable to be matched during one-to-one matching between treatment groups on baseline SBP (±5 mm Hg) using the MATCH procedure,24 and were thus excluded. Matching was performed because the mean baseline SBP among the remaining patients was significantly higher in the OROS-MPH–treated group (117.8 mm Hg) compared with the placebo-treated group (114.6 mm Hg; P=.046). Mean baseline DBP was not significantly different between groups. The ±5 mm Hg level was selected in order to maximize the number of participants and statistical power. Among the 230 analyzed participants (115 randomized to OROS-MPH and 115 randomized to placebo) the mean baseline SBPs for the OROS-MPH–treated group and placebo-treated group were 117.8 mm Hg and 115.4 mm Hg, respectively (P=.13). The trial was 11 weeks, but this secondary analysis was analyzed only through week 10 because data for OROS-MPH pill consumption were collected only through week 10.
During the 11-week active treatment phase, participants were randomized to OROS-MPH or placebo. OROS-MPH was increased during the first 2 study weeks from 18 mg/d to a maximum of 72 mg/d, or the highest dose tolerated. All participants received transdermal nicotine patches and were instructed to wear a 21-mg patch daily beginning on the pre-target quit phase date (study day 27) through week 11.
Dependent Variables Primary outcome measures of SBP, DBP, and HR were obtained at baseline and each study visit using an oscillometric device (Omron-HEM-907XL, Kyoto, Japan) after participants had been sitting at rest for at least 5 minutes. Abnormally high or low BP readings (SBP >140 or <90 mm Hg, DBP >90 or <60 mm Hg) were repeated after 5 minutes and averaged with the first reading. BP monitors were calibrated every 12 months against a mercury manometer via a T tube. Cigarettes per day25,26 and medication and nicotine patch adherence were assessed weekly via self-report and pill/patch count.
Independent Variable and Covariates The primary independent variable was treatment assignment to either OROS-MPH or placebo. Potential covariates of increased BP were identified a priori: weight at baseline, age, sex, race (white, black, Native American, Asian, “other,”“mixed race”), ethnicity (Hispanic, non-Hispanic), cigarettes consumed per day at baseline and throughout the study (time-varying covariate), amount of nicotine patches used, treatment compliance/dose (pills consumed), and baseline hemodynamic measures.
Baseline BP Categories Baseline BP categories were defined using criteria from the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC 7).27 SBP was defined as (1) normal (<120 mm Hg), and (2) prehypertension (120–139 mm Hg). DBP was defined as (1) normal (<80 mm Hg), and (2) prehypertension (80–89 mm Hg). Additional post-hoc analyses were performed, dividing participants into equal-sized categories using tertile baseline SBP and DBP. The cutoffs for SBP were (1) <111 mm Hg, (2) 111 to 122 mm Hg, and (3) 123 to 139 mm Hg. The cutoffs for DBP were (1) <70 mm Hg, (2) 70 to 77 mm Hg, and (3) 78 to 89 mm Hg.
The primary data analyses were mixed linear model analyses of repeated measures conducted on SBP and DBP. Restricted maximum likelihood estimation and type 3 tests of fixed effects were used, with the Kenward-Roger correction28 applied to the first-order autoregressive covariance structure. These covariate-adjusted mixed models evaluated main effects and two- and three-way interaction effects (treatment group, baseline BP category, and time). In primary analyses of SBP and DBP, baseline BP was categorized using JNC 7. The timing of the treatment effect was examined using the treatment group × time period interaction. The least squares means of weekly differences in adjusted estimates of SBP and DBP between treatment groups were produced.
In post hoc mixed model analyses designed to address group-size power differences, baseline BP was categorized into equal-sized tertile groups. Cohen’s d was calculated and interpreted as the effect size estimator for the omnibus between patients Treatment group effect within each baseline BP category. In additional post hoc analyses, treatment effects (OROS-MPH vs placebo) on participants with baseline normal BP were examined using simple logistic regression models to estimate the odds of having prehypertension during two consecutive weeks. The 95% profile likelihood ratio confidence intervals were calculated for the odds ratios and the likelihood ratio chi-square statistic was used to test for a significant association between treatment status and each BP category. Among the 230 participants, 378 of 2070 (18.3%) hemodynamic readings were missing. This study’s PROC MIXED procedures made use of all available data and provided a more robust mechanism (restricted/residual maximum likelihood estimation and first-order autoregressive covariance structure) for handling data that are assumed missing at random.29–31
Statistical analyses were conducted using SAS software, version 9.2 (SAS Institute, Inc, Cary, NC). The level of significance for all tests was set at α=0.05 (two-tailed). This study was approved by the institutional review boards of the participating sites.
Demographics and Baseline Characteristics
Mean age and sex were similar between OROS-MPH and placebo groups (Table I). At baseline, participants had approximately equivalent consumption of cigarettes per day. Most participants had normal SBP (60%) and DBP (78%) at baseline (Table I). At baseline, participants randomized to OROS-MPH (n=115) did not have significantly different SBP (P=.13), DBP (P=.42), or HR (P=.33) compared with participants randomized to placebo (n=115). The number of cigarettes smoked per day during the trial was not significantly different between treatment groups among participants with baseline normal SBP (P=.22) or DBP (P=.16) or baseline prehypertension SBP (P=.96) or DBP (P=.50).
Table I. Participant Demographic and Baseline Characteristics (n=230)a
aMeans and standard deviations. bSeventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC 7).
Systolic blood pressure, mm Hga
Normal (<120 mm Hg)
Prehypertension (120–139 mm Hg)
Low (<111 mm Hg)
Middle (111–122 mm Hg)
High (123–139 mm Hg)
Diastolic blood pressure, mm Hg
Normal (<80 mm Hg)
Prehypertension (80–89 mm Hg)
Low (<70 mm Hg)
Middle (70–77 mm Hg)
High (78–89 mm Hg)
Heart rate, beats per min
Cigarettes per day
Covariate-adjusted mixed models analyses of SBP and DPB (Table II) revealed that treatment with OROS-MPH was associated with a significant increase in SBP (P=.0004) and DBP (P=.010) over the entire study period (Figure 1A and 1B). Thus, OROS-MPH–induced increases in BP were evident even after accounting for baseline differences in BP. Baseline SBP, Hispanic ethnicity, and sex were significantly associated with SBP. Body weight, Hispanic ethnicity and baseline DBP were significantly associated with DBP. The interaction between treatment and baseline DBP was significant (P=.038), and the interaction between treatment and baseline SBP trended towards significance (P=.063). These interactions between treatment and baseline SBP/DBP are depicted in Figure 1. Participants with baseline normal SBP had generally higher increases in adjusted estimates of SBP with OROS-MPH compared with placebo (Figure 1E). The same pattern was evident among participants with baseline normal DBP treatment with OROS-MPH associated with higher adjusted estimates of DBP, compared with placebo (Figure 1F). In contrast, differences in adjusted BP estimates between OROS-MPH and placebo in participants with baseline prehypertension SBP or DBP were less evident (Figure 1C,D). Comparing OROS-MPH treatment to placebo, average weekly adjusted increases in SBP and DBP for participants were 4.7 mm Hg (P<.0001) and 3.6 mm Hg (P<.0001), respectively, for those with baseline normal BP and 1.5 mm Hg (P=.27) and 0.4 mm Hg (P=.79) for those with baseline prehypertension. Effect sizes (OROS-MPH treatment compared with placebo) for baseline normal BP (SBP, d=.73; DBP, d=.76) were more than 3 and 10 times greater than the effect sizes for baseline prehypertension levels (SBP, d=.23; DBP, d=.07), respectively.
Table II. Main Effects and Interaction Effects Associated With Increases in Systolic Blood Pressure and Diastolic Blood Pressure, Weeks 2–10
Systolic Blood Pressure
Diastolic Blood Pressure
Abbreviation: OROS-MPH, osmotic-release oral system methylphenidate. aBlood pressure categories (Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure) for (1) systolic blood pressure: normal (<120 mm Hg), prehypertension (120–139 mm Hg), and (2) diastolic blood pressure: normal (<80 mm Hg), prehypertension (80–89 mm Hg).
Cigarettes per day
Blood pressure categorya
Treatment (OROS-MPH vs placebo)
Number of treatment pills consumed
Number of nicotine patches used
Cigarettes per day
Treatment×blood pressure category
Blood pressure category×time, wk
Blood pressure category×treatment×time, wk
Post hoc mixed models of SBP and DBP were repeated using tertile categorization criteria for baseline BP. Baseline BP was significantly associated with both SBP (P<.0001) and DBP (P<.0001). The interactions between OROS-MPH treatment and baseline BP were significant for both SBP (P=.028) and DBP (P=.007). Lower baseline BP was associated with greater OROS-MPH–associated increases in SBP and DBP. Average weekly adjusted increases in SBP and DBP for the lowest, middle, and highest tertiles were 4.8 (P=.0003) and 5.2 mm Hg (P<.0001), 4.2 (P=.0006) and 2.4 mm Hg (P=.009), and 0.3 (P=.82) and 0.8 mm Hg (P=.44), respectively (Figure 2).
Risk of OROS-MPH–Induced Prehypertension, High-Normal, and Hypertensive BP Values Among Participants With Normal Baseline BP
Treatment with OROS-MPH was significantly associated with increased odds of 2 consecutive weekly prehypertensive BP measurement among participants with normal baseline SBP and DBP, relative to placebo (Table III), during the 10-week study. The predicted odds of 2 consecutive measurements of systolic or diastolic BP in the prehypertension range during weeks 2 through 10 were 3.32 (P=.006) and 4.32 (P=.004) times greater, respectively, among baseline normal participants randomized to OROS-MPH compared with placebo.
Table III. Odds of Two Consecutive Weekly Prehypertensive Measurements Among Participants With Both Baseline Normal SBP (<120 mm Hg) and DBP (<80 mm Hg)a
Abbreviations: DBP, diastolic blood pressure; OROS-MPH, osmotic-release oral system methylphenidate; SBP, systolic blood pressure. aOccurrence of prehypertension during two consecutive weeks. b95% Confidence interval (CI) calculated from profile likelihood. cP values calculated from likelihood ratio. dOccurrence of both systolic and diastolic prehypertension in the same patient for two consecutive weeks; systolic and diastolic prehypertension not necessarily concurrent.
Systolic prehypertension 120–139 mm Hg
Diastolic prehypertension 80–89 mm Hg
Both systolic/diastolic prehypertensiond
Weight Changes During Trial
Follow-up weights were obtained at weeks 6 and 11. Within-treatment group comparisons demonstrated no significant weight loss across follow-up. The same was true for participants with both baseline normal SBP and baseline prehypertensive SBP.
The present study suggests two findings: (1) adult ADHD participants with baseline normal BP are more susceptible to the BP-raising effects of methylphenidate than those with baseline prehypertensive BP, and (2) normotensive adults with ADHD have significantly greater odds of developing prehypertension when treated with OROS-MPH, compared with placebo. The clinical implications of increased susceptibility to stimulant-induced BP increases among persons who are normotensive, compared with prehypertensive, remain unknown. Since previous studies have shown adverse cardiovascular outcomes associated with prehypertension, further studies are needed to clarify the long-term consequences, especially given the growing consensus of ADHD as a treatable lifelong illness.32
Relationship Between Baseline BP and OROS-MPH–Induced BP Elevation
While it is well known that prescription stimulants, such as OROS-MPH, are associated with sustained increases in BP and HR,3–11,33 underlying predictors have not been determined. Wilen and colleagues reported an inverse relationship between baseline BP and HR and ADHD medication-induced hemodynamic measures, noting that “adults with normal BP/HR may be more susceptible to minor cardiovascular changes by sympathomimetics than those with elevated vital signs.”34 However, only 18 participants in the study were treated with MPH and the study included the nonstimulants buproprion and desipramine. Furthermore, the actual risks associated with MPH treatment according to baseline BP were not described. Thus, ours is the first study demonstrating that patients with normal baseline BP may be at heightened risk of stimulant-induced BP increases relative to those with baseline prehypertension.
Stimulants and the Sympathetic Nervous System
Mechanisms underlying MPH-induced increases in BP are unknown but are thought to result from increased release of dopamine in the central nervous system that lead to increased sympathoadrenal stimulation.35 Our study did not directly measure sympathetic activity or plasma catecholamines. However, the differences in BP response we observed in participants with baseline normal and prehypertensive BP may be related to MPH’s effects on the sympathetic nervous system, both peripherally and centrally. For example, sibutramine, an amphetamine-like weight-loss drug withdrawn from the US market due to adverse cardiovascular outcomes,36 reduces muscle sympathetic nerve activity (SNA).37 The sibutramine-induced increase in BP was inversely proportional to the participants’ baseline SNA. When basal nerve traffic was low, sibutramine predominantly inhibited peripheral norepinephrine (NE) reuptake, causing BP elevation. With high baseline SNA, the effects of sibutramine on the central NE transporter predominated, resulting in net reduction of SNA and BP. Whether MPH exerts the same effects on regulation of sympathetic outflow and NE release as sibutramine remains to be investigated.
Additionally, MPH may enhance vascular adrenergic sensitivity, thus augmenting pressor responses in patients with lower baseline BP. Although we did not measure SNA in study participants, a prior study has shown that SNA is lower in normotensive patients than those with “borderline hypertensive” BP.38 Sympathetic overactivity leads to downregulation of vascular alpha-adrenergic receptors, while efferent sympathetic failure (ie, underactivity) leads to denervation supersensitivity. Charkoudian and colleagues39 demonstrated that vasoconstrictor responses to intrabrachial infusion of NE and tyramine are augmented in otherwise healthy patients with low muscle SNA. Alternatively, studies have shown vasoconstrictor responses to NE in hypertensive patients compared with normotensive controls to be either equivalent40 or increased.41–43 Whether ADHD patients with normal BP have low SNA and enhanced vascular adrenergic sensitivity, resulting in increased susceptibility for MPH-induced BP elevation, remains to be further investigated.
Our findings suggest that adults with baseline normal BP may have a significant risk of developing prehypertension and high-normal BP when treated with OROS-MPH. Prehypertension alone has been associated with a 70% increased risk of coronary artery disease and 3.5-fold increased risk of myocardial infarction,44,45 while high-normal BP was associated with a 60% to 80% increased risk of cardiovascular events compared with optimal BP (<120/80 mm Hg based on JNC 6) in the Framingham Study.46 In light of these findings, this study’s results, if confirmed, suggest increased clinical monitoring of all patients taking prescription stimulants and not just patients with higher baseline BP.
It is well known that BP elevations are associated with increased incidences of adverse cardiovascular outcomes, such as myocardial infarction and stroke.12 Because patients with prehypertension are thought to be at greater risk for adverse cardiovascular outcomes compared with normotensive patients, it may be reassuring that BP increases are modest in comparison. On the other hand, it is possible that the relationship between stimulant-induced BP increases and adverse cardiovascular events is not linear. That is, perhaps small increases in BP among the prehypertensive population may have greater clinical impact than larger increases in the normotensive population. Notably, two observational studies in adults have found a safety signal suggesting a possible association between prescription stimulant use and adverse cardiovascular outcomes.14,15 Further studies are needed.
The strengths of this study include a large randomized multisite placebo-controlled clinical trial with weekly monitoring of BP using the same instruments at each site. Group-size differences (ie, power) between baseline normal and prehypertension groups are unlikely to explain our findings: when group sizes were made equivalent by categorizing baseline BP based on tertiles, similar results were obtained and robust differences in effect size were observed between groups. Weight loss is associated with reductions in BP,47 and methylphenidate use is associated with weight loss.6 However, significant weight loss was not observed between and within treatment groups, even when examined by baseline SBP. Thus, OROS-MPH–induced weight loss is unlikely to explain the study results.
Because treatment groups significantly differed in baseline SBP (P<.05), a one-to-one matching strategy on baseline SBP between treatment groups was utilized. We chose the ±5 mm Hg level, as this maximized the power of the study while rendering statistically insignificant the difference in baseline SBP between treatment groups. In a post hoc analysis (results not reported), we also matched at the ±1 mm Hg level, which yielded groups that had virtually identical baseline SBP, but was also much smaller (n=190). The lack of a significant interaction between OROS-MPH treatment and baseline BP in this smaller sample could suggest that such a relationship does not exist, or alternatively, may be a type II error due to lack of statistical power in this smaller sample. Replication of the study in a different population sample is therefore required. It should be noted that baseline DBP was not significantly different in the study population, and the same general findings were observed for DBP as they were for SBP.
Further limitations of this study include (1) the study was not designed/powered to examine the changes in BP and HR associated with OROS-MPH, (2) it is unknown whether the results can be extrapolated to other prescription stimulants, (3) results may not be generalizable to nonsmokers or adults without ADHD (however, adults with ADHD have a much higher prevalence of cigarette use than the general population21), (4) and cigarette use treatment with nicotine patches could affect hemodynamic values (all patients received patches, which have minimal hemodynamic effects on persons with nicotine dependence,48,49 and cigarettes smoked per day were similar between participants with baseline normal and prehypertension BP), and (5) potentially important unmeasured covariates such as cardiovascular family history could not be included, but this potential bias is mitigated by treatment randomization. Regression to the mean refers to how in cases of outlying initial measurements, subsequent measurements may be nearer group means. Regarding the issue of regression to the mean in this study, concerns are mitigated given that BP increased in both placebo and OROS-MPH–treated groups among participants with lower baseline BP, but BP did not decrease in participants with higher baseline BP levels. Moreover, regression to the mean is unlikely to explain the differential treatment effect observed, given that regression to the mean would be expected to apply to the entire study population and not differentially between treatment groups (ie, outlying initial measurements in all groups expected to be nearer group means in subsequent measurements).
This study of adult smokers with ADHD suggests a greater OROS-MPH–induced BP elevation in participants with normal baseline BP compared with those with prehypertension at baseline. The finding of more modest OROS-MPH–induced BP increases among those with prehypertension at baseline is of uncertain clinical significance. This study also calls attention to the potential risks of prescription stimulant use inducing prehypertension in normotensive adults. Differences in baseline SBP among treatment groups, in particular, makes replication of this study necessary.
Disclosures: Arthur Westover has served as an expert witness to a private university. Bryon Adinoff has served as a consultant for Shook, Hardy & Bacon LLP (medical malpractice consultant, tobacco companies), GlaxoSmithKline, and Teva Pharmaceutical Industries Ltd. The other authors have no conflicts of interest or financial disclosures to report. This clinical trial was registered at http://clinicaltrials.gov (identifier: NCT00253747).
Sources of Funding: Funding to support this study was received from NIH/NIDA Clinical Trials Network: North Texas Node (3U10DA020024-01S1), NIH CTSA Grant UL1RR024982 and the O’Brien Kidney Center. These funding organizations had no role in the design and conduct of the study or preparation of the manuscript.