Author + information
- Kirk N. Garratt, MSc, MD∗ ()
- Department of Cardiovascular Medicine, Lenox Hill Hospital, North Shore/LIJ Health System, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. Kirk N. Garratt, Lenox Hill Hospital, 130 East 77th Street, 9th floor, New York, New York 10075.
As an interventional fellow, I hated to miss cases. I came in early and stayed late. I took call without hesitation. I was always available, and I never let lack of sleep get in my way. I was indefatigable—it was a form of invincibility. I’d remind myself of that as I worked exhausting hours, often up most or all of the night, unwilling to accept fatigue as a reason to quit.
The peril of such an attitude, considered central to a good work ethic a few years ago, seems obvious. Studies document what we all know intuitively: acute and chronic sleep deprivation slows response time, impairs fine hand–eye coordination, and affects judgment (1)—elements key to a safe and effective angioplasty. The biology of acute sleep deprivation is well studied, with some work showing measureable impairment after as little as 17-h continuous wakefulness (2); paradoxically, prolonged wakefulness can impair insight such that one feels less vulnerable, not more, to making errors. Sleep deficiency has been documented to lower physician performance in some surgical and technical disciplines, but not all (2,3). We’ve learned hard lessons about the risks of residents persisting when exhausted through the Libby Zion and other disastrous cases that led to federally mandated work-hour rules for medical trainees. Regulators and others focused on patient safety have also suggested restricting the work hours of practicing physicians to minimize preventable patient injury (4,5). Is it time to regulate the work hours of interventional cardiologists?
In this issue of JACC: Cardiovascular Interventions, Aronow et al. (6) considered short-term patient outcomes when percutaneous coronary intervention (PCI) was performed the day after a call night during which at least 1 PCI was performed between midnight and 6 am. They wanted to see whether short-term sleep deprivation influenced in-hospital mortality, bleeding complications, and other composite metrics. Adjusted outcomes for patients treated by short-term sleep-deprived operators were compared with outcomes for patients treated by operators without known sleep deprivation using propensity analysis and 2:1 greedy matching. The impact of more long-term sleep deprivation, assumed to exist if an operator conducted >1 middle-of-night PCI within 7 days, was assessed through similar methods. National Cardiovascular Data Registry (NCDR) CathPCI Registry data were used, which provided a huge sample size. Their key observations were:
1. No relationship was found between short-term sleep deprivation and any outcome metric.
2. Later start times for day-after cases, assumed to involve the most fatigued operators, did not correlate with increased risk.
3. Long-term sleep deprivation was linked to increased risk of bleeding complications, but not other metrics of performance.
4. Daytime PCI after a busy night on call was uncommon, accounting for <1 in 40 PCIs registered in the NCDR CathPCI Registry.
The ascendency of a safety culture in American medicine is allowing civil conversations to take place around thorny topics, including regulation of physician work hours to assure patient safety. The current paper by Aronow et al. (6) represents the largest published analysis of interventional cardiologist fatigue and patient safety, and its findings corroborate similar findings reported by Crudu et al. (3), in the only other published (but much smaller) study specific to interventional cardiology. In short, despite good reasons for expecting trouble with short-term sleep deprivation, none was found. More long-term sleep deprivation had almost as little impact on outcomes. We must conclude that policies restricting interventionalist access to the cath lab the day after call are not justified on the basis of safety-related evidence. So is this large-scale, well-structured, contemporary study the final word?
The NCDR CathPCI Registry provides enormous power in making certain observations because of its size, but registry analyses are notoriously limited. NCDR captured how late the day-after PCIs were done, but didn’t capture which physicians got some rest, or how much. Because some practices encourage doctors to rest and come in later if they’ve been up most of the night, the authors’ assumption that later PCI start times for day-after cases reflected work by the most fatigued doctors might be wrong. We don’t know how many of these physicians had a disrupted night of sleep as opposed to a completely sleepless night. Also, finding that interventionalists performed well collectively after a night on call doesn’t mean that all physicians performed well, because genetic and other factors are known to affect tolerance of sleep deficiency (7). We don’t know how many physicians regulated themselves by rescheduling cases because they were tired. And of course, NCDR can’t tell us what it was like to deal with tired doctors the day after call. Fatigue lowers the threshold for abrupt, confrontational, or abusive behavior (8); the Joint Commission believes such behavior “undermine(s) a culture of safety” (9), and it certainly can color patients’ impressions about their care. Fatigue may well have adverse effects on the patient experience beyond what’s measured by NCDR. Importantly, though, NCDR was able to tell us that if any problem exists, its magnitude is small: Very few day-after cases are actually done by short-term sleep-deprived operators.
The only significant finding was the roughly 20% increase in adjusted bleeding rates among patients treated by long-term sleep-deprived physicians. I struggle to understand why bleeding avoidance strategies (use of bivalirudin, radial access, and vascular closure devices ), which were used only slightly less often by sleep-deprived physicians, should be so much less effective in this setting. In many hospitals, other team members are responsible for access management after PCI (usually fellows, nurses, or technicians). Physicians may recover fully from a working call night or night shift after just 2 days of normal sleep (11), so 2 call nights over 7 days, the definition used in the current report, seems to reflect relatively mild long-term sleep deprivation. And in the study by Crudu et al. (3), bleeding complications were actually less common among sleep-deprived physicians. Better understanding of the specific practice elements that affected bleeding would be helpful, but again, the limitations of a registry study are evident.
All this leads to an interesting question: if these findings are correct, why doesn’t short-term sleep deprivation affect interventionalist performance? Are we somehow adapted to disrupted sleep such that we can manage the neurocognitive impact? Is sleep deprivation so pervasive that our baseline performance (the non–sleep-deprived cohort in this study) reflects chronic fatigue, making differences immeasurable? Does a well-rested team make up for a dead-tired doctor? Has angioplasty become so well developed that the safety margin is actually bigger than we imagine? Answers are unknown, but sleep research provides some clues: newly learned skills are most vulnerable to the effects of sleep deficiency, whereas the accuracy of better-established skills may be maintained, albeit at the expense of efficiency (12); perhaps the repetitive habits of angioplasty aid in skill preservation. Also, whereas psychomotor skill erosion from sleep deprivation matches that caused by alcohol intoxication between 18 and 24 h of wakefulness, ability seems to recover thereafter, perhaps related to realignment with normal circadian rhythms (2).
The current analysis may not be the last word, but this work will stand as an important reference piece on the effects of late night work and next-day interventional performance, and should inform discussions about cath lab work-hours policies. Although seemingly contrary to intuition and the tenets of sleep science, no signals of compelling safety concerns for short-term sleep-deprived interventional cardiologists have been found for the hard endpoints available in the NCDR CathPCI Registry. Looking deeper into the totality of patient care, with more granular assessment of the impact of physician and practice specifics, should be undertaken before accepting broad policies that limit physician activities in the catheterization laboratory. This conclusion will likely be challenged by interested groups, such as the Sleep Research Society (SRS), which has proposed legislation requiring a separate informed consent process when an operator has been awake for 22 consecutive hours, and the American Academy of Sleep Medicine, which (along with SRS) has endorsed proposed legislation to define an automobile driver as impaired after a similar period of wakefulness (13). Although these seem sensible and well meaning, no policy or legislation that restricts physician privileges can be justified without evidence of merit, and in the case of day-after angioplasty, the evidence is simply not there.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Garratt is a consultant for Boston Scientific, The Medicines Company, and Abbott Vascular; receives research support from Boston Scientific, CeloNova, and Mayo Clinic Foundation; and has equity in Infarct Reduction Technologies and Guided Delivery Systems.
- American College of Cardiology Foundation
- Pilcher J.J.,
- Huffcutt A.I.
- Aronow H.D.,
- Gurm H.S.,
- Blankenship J.C.,
- et al.