Date: August 14, 2023

Reference: Ma et. al. The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet 2023

Guest Skeptic: Dr. Mike Pallaci is a Professor of Emergency Medicine (EM) for the Northeast Ohio Medical University, and an Adjunct Clinical Professor of EM for the Ohio University Heritage College of Osteopathic Medicine. He has been program director for two EM residency programs, and is currently a core faculty member for the EM residency at Summa Health System in Akron, OH, where he also serves as the Medical Director of the Virtual Care Simulation Lab.

This episode originated because of a thread Mike posted on the social media site formerly known as Twitter.  The tweet said: “I am sick and tired of some non-EM docs/specialists slamming EM when we don’t aggressively lower BP in ICH.” It certainly got a lot of attention.

It got a lot of attention both on the positive side (mostly from EM docs who share the frustration) and on the negative side from some neurologists who didn’t seem to particularly care for the premise of the tweet or for the generally positive response. It started out with venting on Twitter about an unpleasant interaction with one of Mike’s partners and turned into a week-long discussion that culminated in an invitation to be the guest skeptic on the SGEM to critically appraise INTERACT-3.

Mike backed up his position on blood pressure (BP) lowering using evidence. Specifically, he pointed out that the evidence behind the guidelines re BP management in intracranial hemorrhage (ICH) is relatively weak.

This received a mixed response on Twitter. Most EM physicians shared the frustration and made positive comments. Some neurologists disagreed with the premise of the tweet and challenged it.

I responded by posting the 2013 SGEM episode on INTERACT-2 (SGEM#73) which showed no statistical difference between intensive (<140 mmHg) and guideline directed (<180 mmHg).  In 2017, the SGEM reviewed the ATACH-2 trial (SGEM#172) which showed similar results.

A meme was also posted of Charlton Heston from the classic movie the Ten Commandments. It emphasized that GUIDElines are not GODlines. The literature should inform and guide our care, but it should not dictate our care. This is a core principle of evidence-based medicine (EBM). Often the available evidence on a specific medical question is weak. We still need to apply our clinical judgement and ask the patient about their values and preferences.

Case: A 67-year-old male presents to the emergency department obtunded with left hemiplegia.  Symptoms began just prior to presentation.  His blood pressure (BP) is 194/110 mmHg.  CT reveals a hemorrhage in the right internal capsule, suggestive of acute hypertensive hemorrhagic stroke.  Should the blood pressure be treated?  If yes, what should the target blood pressure be? How quickly do we want to get there? And are there any other physiologic variables we want to be aggressive about controlling in the early treatment window?

Background: We have covered the common issue of elevated BP after ICH on SGEM#73 and SGEM#172. The 2022 AHA/ASA guidelines give several recommendations on this topic. The class (strength) of their recommendation is 2a/2b based upon Level B and Level C quality of evidence.


It is really important to pay attention to the specific language used in the guidelines. First of all, as we’ve already covered, a guideline is something developed by humans giving their best interpretation of the evidence to serve as a guide, not something given to Moses on Mount Sinai. But that point aside, the basis of this discussion was that the strength of the evidence behind these guidelines and the strength of the recommendations in the guidelines themselves are frequently misunderstood and/or misrepresented by our consultants. The specific language used in the guidelines is as follows:

  • “In patients with spontaneous ICH in whom acute BP lowering is considered, initiating treatment within 2 hours and reaching target within 1 hour can be beneficial to reduce the risk of HE and improve functional outcome”

They give this a Class 2a recommendation, “Moderate”: is reasonable, can be beneficial; Level of Evidence C-LD (Limited Data). We would say than it would also be reasonable not to acutely lower the BP

  • “In patients with spontaneous ICH of mild to moderate severity presenting with SBP between 150 and 220 mmHg, acute lowering of SBP to a target of 140 mmHg with the goal of maintaining in the range of 130 to 150 mmHg is safe and may be reasonable for improving functional outcomes.”

This was a Class 2b recommendation, “Weak”: may be reasonable, may be beneficial, effectiveness not well established; Level of Evidence B-R (Randomized; moderate quality evidence from 1 or more RCTs or meta-analyses of moderate-quality RCTs). Listeners to the SGEM know when the word “may” is used it can be substituted with the phrase “may not”

  • “In patients with spontaneous ICH presenting with large or severe ICH or those requiring surgical decompression, the safety and efficacy of intensive BP lowering are not well established.”

Another Class 2b recommendation, “Weak”: may be reasonable, may be beneficial, effectiveness not well established; Level of Evidence C-LD (Limited Data).

This does not sound like language describing an intervention that is a “central component of management”. Yet that is what is what many physicians believe the guideline tells us. All you have to do is read the guideline itself to recognize that even the people writing these guidelines, who I would argue are biased toward aggressively treating BP in this scenario, recognize that the answer here is not clear. But when our consultants come to the ED they frequently act as though it is a clear and accepted truth and that this is what we should be doing.

Clinical Question: Can the implementation of a goal-directed care bundle incorporating protocols for early intensive blood pressure lowering and management algorithms for hyperglycemia, pyrexia, and abnormal anticoagulation, implemented in a hospital setting, improve outcomes for patients with acute spontaneous intracerebral hemorrhage?

Reference: Ma et. al. The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet 2023

  • Population: Adults patients 18 years of age and older presenting within 6 hours after the onset of ICH
    • Exclusions: Definite evidence that the ICH is secondary to a structural abnormality in the brain or previous thrombolysis. Also, if the attending clinician felt there was a high likelihood that the patient will not adhere to the study treatment and follow-up regimen.
  • Intervention: A goal-directed intensive care bundle protocol correcting hypertension, hyperglycemia, pyrexia and hypercoagulability, with the goal of achieving treatment targets within one hour of initiating treatment and maintaining them for 7 days (or until discharge or death, whichever came first)
  • Comparison: Usual care at the discretion of the treating physician
  • Outcome:
    • Primary Outcome: Functional recovery measured at 6 months according to the modified Rankin Scale (mRS) score and analyzed as an ordinal outcome (shift across all categories)
      • The authors state that one of the “efforts to improve the success of randomized controlled trials in identifying an effective treatment for ICH have included…to extend the assessment of functional outcome beyond the conventional 90 days, because recovery from ICH takes longer than from acute ischemic stroke.”
    • Secondary Outcomes:
      • Functional recovery according to a shift analysis of scores on the National Institutes of Health Stroke Scale (NIHSS) at seven days.
      • Dichotomous mRS outcomes at 6 months
        • 0-2 vs. 3-6 and 0-2 vs. 3-5 (i.e. – major disability in survivors)
      • Death at 6 months
      • Death or neurological deterioration at seven days
      • Health-related quality of life (HRQoL) using the EuroQoL Group 5-Dimension self-report questionnaire (EQ-5D)
      • Residence at six months (own home vs. other)
      • Time to hospital discharge
    • Safety Outcomes: All-cause and cause-specific serious adverse events were recorded for the duration of follow-up.
  • Type of Study: A pragmatic, international (10 countries), multicenter (121 hospitals), blinded endpoint, stepped wedge cluster randomized controlled trial.

Authors’ Conclusions: Implementation of a care bundle protocol for intensive blood pressure lowering and other management algorithms for physiological control within several hours of the onset of symptoms resulted in improved functional outcome for patients with acute intracerebral haemorrhage. Hospitals should incorporate this approach into clinical practice as part of active management for this serious condition.

Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. Yes
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were randomized. Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Unsure
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  8. All groups were treated equally except for the intervention. Yes
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. Unsure
  12. Financial Conflicts of Interest. The authors state the funders of the study had no role in the design, collection/analysis/interpretation of data, or writing of the report. Only three of 42 authors reported competing interests, two of which (the second author and the last/senior author) were related to a pharmaceutical company.

Results: They recruited a total of 7,036 patients from 121 hospitals that could be included in the modified intention-to-treat analysis. The mean age of patients was 62 years with 36% female. Most of the patients (>90%) were Chinese.

Key Result: The odds of a poor functional outcome was lower in the care bundle group compared to usual care.

  • Primary Outcome: mRS favored the care bundle group (OR 0.86; 95% CI 0.76-0.97, p=0.015), consistent across all adjustments/calculations.
  • Secondary Outcomes: Most secondary outcomes did not show a statistically significant difference. Some showed trends in a positive direction. Patients who received the intervention were statistically more likely to be discharged by day 7. The EQ-5D quality of life assessment was a mixed bag, but the effects on this scale diminished when they made the various statistical adjustments in their post-hoc analysis.
    • Functional recovery according to a shift analysis of scores on the National Institutes of Health Stroke Scale (NIHSS) at seven days – No statistical difference
    • Dichotomous mRS outcomes at 6 months – No statistical difference
    • Death at 6 months – No statistical difference
    • Health-related quality of life – Mixed bag that diminished with adjustments
    • Residence at six months (own home vs. other) – No statistical difference
    • Time to hospital discharge – More likely to be discharged by day 7
  • Safety Outcomes: Significantly fewer serious adverse events in bundled care group

1. External Validity: There are several issues to unpack including the patient population, medications used to lower the BP and the prevalence of patients being anticoagulated.

  • Setting: They included nine low- and middle-income countries (Brazil, China, India, Mexico, Nigeria, Pakistan, Peru, Sri Lanka, Vietnam) and one “high-income country” Most of the cohort came from China (90%) with only 3% coming from Chile.
  • Medications: The medications used to lower blood pressure included urapidil 61%, nipride 21%, labetalol 12%, nicardipine 8%, nimodipine 8%. How would this data extrapolate to the North America where I’m guessing nicardipine would likely be ~80% and urapidil wouldn’t even be on the radar as it not FDA or Health Canada approved.
  • Anticoagulation: ~1% of the patients in the cohort were anticoagulated. I suspect that this would be higher in North America. The 2020 census of USA reported 258 million adults. More than 8 million are on anticoagulants (3%). Since being on blood thinners increases the chance of ICH that percentage would be higher. This higher percentage could impact the efficacy of lowering the blood pressure with patients on Vitamin K antagonists and direct oral anticoagulants (DOACs)

2. Bundle: The bundle treatment included addressing the hypertension, hyperglycemia, pyrexia and hypercoagulability Which part of the intervention caused the benefit?

  • ~90% had BP elevation, ~1/3 hyperglycemia; other markers ~1-2%. It would be reasonable to assume BP would be the primary driver of the results, but we do not know
  • We have seen with other “bundles” like sepsis that did not ultimately turnout to be better than usually care (ARISE, ProMISe and ProCESS)
  • They plan to do a mediation analysis to identify the effects of the various components of the bundle
  • There could also be the Hawthorne Effect because this was unblinded and participants and clinicians knew whether they were doing bundled care or usual care

3. Outcome Assessment: The primary outcome in this trial was an ordinal analysis of the mRS score. Concerns have been raised about ordinal analyses. See Dr. Rory Spiegel’s references below.

There are also the issues with the mRS score itself for inter-rater reliability (IRR).  Scoring of mRS, even by a neurologist, is only moderately reliable at best when done face to face. Quinn et al 2009 reviewed 10 studies and report the overall reliability of the mRS was kappa 0.46 (moderate) and conclude: There remains uncertainty regarding modified Rankin Scale reliability.” Wilson et al 2005 found similar issues with the mRS.

Neurologist Dr. Ravi Garg, highlighted that the IRR of the mRS was lowest at the junction of mRS 0-1 (good outcome) and mRS 2-6 (bad outcome), leading to what he calls “endpoint wobble.”  I suspect the wobble is even greater in this trial because the outcome assessment is done via caregiver over the phone (79% of the time) or with the patient over the phone (7% of the time) and not face to face (<1% of the time).

4. Strength of the Evidence: The people writing these papers and many in clinical practice seem to continue to overstate the strength of evidence for lowering BP in ICH. It is in the introduction and discussion in INTERACT-3. Again, the evidence ranges from weak to flawed to non-existent, and the language of the guidelines reflect the strength of the evidence. However, the language used by the INTERACT-3 authors and many of our consultants seems to be much stronger than what is justified by the guidelines and the data that they are based on.

5. Refused or Lost to Follow-Up: There was 2.6% of the cohort that had missing outcome data. In addition, 8.5% of patients either refused (1%) or were lost (7.5%) to follow-up. It was similar between both groups. However, the absolute difference between bundle and usual care for mRS (0-2) was only 3.6%.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We disagree with the authors that hospitals should incorporate this care bundle protocol that includes intensive blood pressure lowering into the approach into their clinical practice. The missing data in addition to the lack external validity, use an unmasked bundled treatment and reliance on phone assessments using a mRS score with at best moderate IRR all leads to more uncertainty about the results. If they had concluded that, based on their findings, implementation of the bundle should be considered, particularly in settings with limited resources and/or access to care, we might be able to get on board.

SGEM Bottom Line: Intense blood pressure lowering in patients with intracranial hemorrhages continues to seem safe as long as we don’t overshoot our target BP and cause hypotension, but we are still uncertain if it provides a patient-oriented outcome of benefit.

Case Resolution: Our patient has a terrible disease, severe symptoms, and a marker of disease severity (severe hypertension) that makes his prognosis poor. Unfortunately, despite what some of our consultants may tell us, there is no definitive evidence to inform our management. You have a conversation with the neuro critical care team who will be taking care of the patient in the hospital and work out a treatment plan that is mutually agreeable.

Dr. Mike Pallaci

Clinical Application: The results of INTERACT3 and prior studies provide fairly strong evidence that treating the blood pressure in this scenario is safe, and that is reflected in the AHA/ASA guidelines.  However, multiple trials (including this one) have shown either weak proof or no proof of benefit. The 2022 AHA/ASA ICH Guidelines state that aggressive treatment is “safe,” and “may be effective” for patients with mild to moderate ICH, but “not well established” in patients like ours with severe ICH; they also state that acute lowering of SBP to <130 mmHg “is potentially harmful.”  This language leaves plenty of room for clinical judgment, and the questions surrounding the “positive” trials in addition to the publication of other “negative” trials including ATACH2 make a conservative approach to blood pressure management, at minimum, acceptable.

Dr. Gary Hunter

Dr. Gary Hunter is an Associate Professor and Division Head of Neurology at the University of Saskatchewan. He was part of the original Twitter discussion on this topic. Gary was invited to be a guest on this SGEM episode and comment on the Twitter exchange, how he would resolve the case presented and give his expert opinion on how to clinically apply INTERACT-3. This was offered as an opportunity for discussing the nuance of the literature, common ground in caring for patients and that the evidence can be interpreted and applied differently. Gary was away on a family vacation and not able to participate on this SGEM episode but will comment on the episode when possible. We respect that family comes first and look forward to hearing Gary’s views when he is back from holidays.

What Do I Tell the Patient?  As always…it all depends. This is where we earn our money, making tough decisions with inconclusive evidence while taking multiple factors into account. I tell the patient (or in this case, his family) that there are some who believe that being aggressive about managing the blood pressure is beneficial, but that I’m not at all convinced by the evidence, and neither are those who write the guidelines. One thing we can all agree on is that overshooting our target and causing hypotension is potentially harmful. Without definitive evidence to inform our decision, we’ll come to the best conclusion and treatment recommendation that we can.

Keener Kontest: Last weeks’ winner was repeat win for Scott Luce. He knew George Cottrell invented the cervical spine collar.

Listen to the SGEM podcast for this weeks’ question. If you know the answer, then send an email to with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics Guide to Emergency Medicine.

Additional Reading: