Presence of staff does not affect CVD risk reduction in intensively treated SPRINT group
Blood Pressure Measurement in the Systolic Blood Pressure Intervention Trial (SPRINT)
Presented at AHA 2017 Scientific Sessions in Anaheim, CA, USA by Karen C. Johnson (University of Tennessee Health Science Center, Memphis, TN, US)Nov. 14, 2017 - news
The aim of the SPRINT trial was to examine the effect of a more intensive blood pressure (BP) treatment than currently used by randomizing patients to intensive treatment with a target value <120 mmHg or standard treatment aimed at <140 mmHg. The primary endpoint was a composition of MI, ACS, stroke, chronic HF or CVD mortality.
BP measurement in the SPRINT study corresponded to the method as used in almost all recent HTN outcomes studies, and as recommended for clinical practice by virtually all guidelines. The measurement procedure stated that the BP was the average of three measurements with an automated device after a 5 minute rest period. SPRINT results showed that treatment to intensive target values lowered the risk of CVD and mortality compared to standard treatment. There has been a lot of discussion among researchers not involved in the SPRINT study that the measurement was done in the absence of the physician, which meant that the measured BP was not comparable to that in other studies. Johnson stressed that the SPRINT study did not investigate the impact of staff presence, nor has it been published so far.
This presentation concerned the results of a survey taken immediately after the last study visits, with questions about whether BP measurements were usually conducted in the presence or absence of staff. The goal of this survey was to examine whether variation in BP measurement method was associated with differences in acquired BP, use of medication, CVD or mortality outcomes or serious adverse events (SAE). There were four BP measurement categories:
- always alone (AA: 4082 participants at 38 sites): participant was alone during 5 minute rest period and the 3 measurements
- never alone (NA: 2247 participants at 25 sites): personnel in the room during BP measurements
- alone for rest period (AR: 1746 participants at 19 sites): participant was alone during rest period
- alone for BP measurements (ABM: 570 participants at 6 sites): participant alone during BP measurements
- There were some differences in baseline characteristics in the four measurement groups.
- In all groups, a SBP of 120 mmHg was obtained in the intensive group, and a SBP of 135 mmHg in the standard group. The difference in SBP between the two treatment arms ranged from 12.5 to 14.7 in the four measurement groups.
- Heterogeneity in the treatment effect of the primary endpoint was seen between the measurement groups. In the AA and NA groups, the intensive treatment group had a lower risk at the primary endpoint than the standard treatment group, whereas in the smaller AR and ABM categories no difference was seen in treatment effect between the study arms. When AA was compared to NA, no interaction was seen with the treatment effect.
- There was no difference in total number of SAEs between the intensive and standard groups in each of the four measurement groups and no heterogeneity in treatment efficacy was seen.
One of the limitations of this analysis is the design of a post-hoc survey, which is based on the memory of the researchers involved, so there is a risk of misclassification. Measurements in the presence and absence of a staff member were not performed in the same patients. In addition, this survey was not specified in advance. Nevertheless, the SPRINT team concludes that similar BP levels and CVD risk reduction were seen in the intensive group, regardless of whether BP measurement was done in the presence or absence of a staff member. In order to fully utilize the benefits and to minimize the risks associated with following the SPRINT intensive treatment algorithm, a validated automated BP device, trained staff that allows a quiet rest period, good positioning of the patient, using proper cuff-size, and averaging of multiple measurements seem to be more important than whether the BP measurement is done in the presence or absence of staff. More research is needed, with better methods to draw stronger conclusions, to determine whether presence or other factors during BP measurement determine the outcome.
Discussant Sripal Bangalore (NYU Langone Medical Center, New York City, NY, USA) had some comments on this presentation. He accepted that aiming for an intensive BP target reduces the risk of CV events compared to a standard target. He questions the precise BP target though. Reason for this unanswered question is the debate after the publication of the SPRINT results on whether the BP measurements were performed in the presence or absence of staff. The method section of the SPRINT publication reported, "the staff member should leave the room during this 5 minute rest period".
Various studies have suggested that the method of measuring influences the measured BP value. Lower values were obtained for instance with automated measurement vs. manual auscultative BP measurement. BP has been shown to rise when a nurse enters the room, and the increase is even greater when it concerns a doctor. Even automatic measurements are higher in the presence of a physician.
In this analysis of SPRINT was, by lack of patient level data, asked per study site how BP measurement was performed. This reveals a lot of variation, despite the protocol, but in almost ¾ of sites nobody was in the room for at least some of the time during measurement. Because this information was gathered based on recall and each site was categorized in one of the four categories, there is risk of misclassification. Uncertainty or inaccuracy can have both small or large consequences for interpretation of the data.
Bangalore calls for being cautious when concluding anything about whether attended or unattended BP measurement affects the measured SBP, as these are indirect comparisons. He notes that it is hardly surprising that SBP was comparable in the four measurement groups, since maintaining a stable SBP was the study aim. The primary endpoint and all-cause mortality were also comparable, which does not surprise him either, because the HR is dependent on the BP-difference between the intensive and the standard treatment groups, not on the absolute BP values that are the topic of debate. He thinks it is more important that the absolute rates of serious adverse events in the intensively treated arm were also comparable among groups, but again there might be confounding at stake due to the misclassification bias.
In conclusion he says that the heterogeneity of method of measurement is striking, considering the instructions and training given. The results are somewhat reassuring, but there are considerable limitations and unanswered questions, among which recall bias, misclassification bias and extrapolation of site-level data to analysis at the patient level. The difference in measured BP values between correct and incorrect measurements is patient-dependent and therefore not a matter of correcting by a set value. Given that the BP values from manual auscultatory measurement does not equal AOP, nor ABP, it is not prudent to have the same BP target for different BP measurement techniques.
- Our reporting is based on the information provided at the AHA 2017 congress -