Letters
COMMENT & RESPONSE Electrocardiographic artifacts can simulate native elec-
Uncovering the Truth of Electromechanical
Association Artifact With Limb Lead Electrodes
To the Editor Drs Zhai and He presented an interesting and well
documented case of a person in their 50s with persistent chest
pain, ST-segment elevation, and QTc interval prolongation.1
The authors provided a thorough explanation of the electro-
cardiographic mechanism of ST elevation, which was caused
by the artifact of right radial artery pulsation. However, this
article points out that “the artifact due to the electrode’s in-
termittent contact with the patient’s skin during systolic pul-
sation of the radial artery was recorded on the ECG.” This find-
ing, referred to as an arterial pulse-tapping artifact, is generated
by the movement of the electrode with each pulsatile motion
of blood flow through the right radial arterial. A pulse-
tapping artifact demonstrates an electromechanical associa-
tion with the QRS complex and a fixed coupling-interval after
the QRS complex, which represents the time delay between car-
diac systole and peripheral pulse perfusion. Therefore, it is ob-
vious that the right radial artery raises the clip-cuff electrode
in diastole rather than systole, resulting in electrocardiogram
artifact.
Electromechanical association artifact is an electrocardio-
gram artifact in which an arterial pulsation distorts the under-
lying electrocardiogram waveforms. Because it is “heart-
made,” it is synchronous with underlying rhythm and does not
separate with the native electrocardiographic waveforms—
generally expected from artifacts because their cycle length
and the underlying rhythm are usually different. This asso-
ciation causes a consistent distortion of the underlying wave
pattern and can mimic myocardial infarction.2
Yun-Tao Zhao
Ya-Lei Han
Author Affiliations: Department of Cardiology, Aerospace Center Hospital,
Beijing, China.
Corresponding Author: Ya-Lei Han, Department of Cardiology, Aerospace
Center Hospital, 15 Yuquan Rd, Beijing 100049, China (10689714@
bjmu.edu.cn).
Published Online: February 7, 2022. doi:10.1001/jamainternmed.2021.8298
Conflict of Interest Disclosures: None reported.
1. Zhai HL, He Y. Transient ST-segment elevation and QTc interval prolongation
in a patient with persistent chest pain. JAMA Intern Med. 2021;181(12):1652-1653.
doi:10.1001/jamainternmed.2021.6015
2. Aslanger E, Yalin K. Electromechanical association: a subtle
electrocardiogram artifact. J Electrocardiol. 2012;45(1):15-17. doi:10.1016/
j.jelectrocard.2010.12.162
In Reply We appreciate the insightful comments by Drs Zhao
and Han in response to our contribution to the Challenges in
Clinical Electrocardiography.1 They raised a concern and we
sincerely apologize for the incorrect wording, ie, “the artifact
due to the electrode’s intermittent contact with the patient’s
skin during systolic pulsation of the radial artery was re-
corded on the ECG.” In fact, artifacts should be generated dur-
ing diastolic rather than systolic pulsation, as Drs Zhao and Han
state.
jamainternalmedicine.com
© 2022 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a Universidad de Concepcion (UDEC) User on 11/28/2022
trocardiogram waves. Artifactual deflections may simulate the
appearance of ventricular fibrillation, ventricular tachycar-
dia, or ST-segment shift.2 In electrocardiograms, these mor-
phologically abnormal ST segments and T waves are called elec-
tromechanical association (EMA) artifacts. An EMA artifact is
a “heart-made” artifact caused by arterial or precordial pul-
sations at the site where the limb or chest leads are placed. In
artifacts generated by the heart, artifactual waves synchro-
nize with the cardiac rhythm.3 These EMA artifacts are the re-
sult of the arterial pulse being transmitted to the lead clip,
which generates the artifact. Modern devices record only leads
I and II and deduce the waveforms of other island chains from
these 2 leads.4 Therefore, most limb leads will be affected.
However, according to the limb leads that generate artifacts,
usually 1 lead is normal in EMA artifacts. This is the most im-
portant clue for diagnosing EMA artifacts.
Yanhong He, MD
Author Affiliation: Department of Cardiology, The Third People’s Hospital of
Yancheng City, Yancheng, Jiangsu, China.
Corresponding Author: Yanhong He, MD, Department of Cardiology, The Third
People’s Hospital of Yancheng City, Xindu Road, Yancheng, Jiangsu 224000,
China (heyanhongdiyi@yeah.net).
Published Online: February 7, 2022. doi:10.1001/jamainternmed.2021.8295
Conflict of Interest Disclosures: None reported.
1. Zhai HL, He Y. Transient ST-segment elevation and QTc interval prolongation
in a patient with persistent chest pain. JAMA Intern Med. 2021;181(12):1652-1653.
doi:10.1001/jamainternmed.2021.6015
2. Aslanger E, Yalin K. Electromechanical association: a subtle
electrocardiogram artifact. J Electrocardiol. 2012;45(1):15-17. doi:10.1016/
j.jelectrocard.2010.12.162
3. Rajendran G, Muthanikkatt AM, Nathan B. Not all waves are factual. Circulation.
2021;144(9):751-753. doi:10.1161/CIRCULATIONAHA.121.055522
4. Aslanger E. Maybe a dazzle but not puzzle. J Electrocardiol. 2010;43(6):682-
684. doi:10.1016/j.jelectrocard.2010.04.010
For and Against Routine Removal of Peripheral
Intravenous Catheters
To the Editor Dr Buetti and colleagues reported that replace-
ment of peripheral intravenous catheters (PVCs) only when
clinically indicated led to an increased risk of bloodstream in-
fection (BSI) compared with routine replacement every 96
hours.1 Overall, their conclusion was supported by the data;
however, at least 3 questions remain.
First, despite an impressive incidence rate ratio of 7.20
(95% CI, 3.65-14.22), a complementary absolute metric that
paints a complete picture is missing. We extracted data from
eFigure 2 in the Supplement1 to evaluate the dwell time per
PVC (dividing PVC days per month by the number of PVCs in-
serted per month). The mean dwell time across all patients was
3.66 days. The absolute increase in BSI incidence rate after the
policy change was 0.77 per 10 000 PVC days (0.9 per 10 000
in the intervention group minus 0.13 per 10 000 in the base-
line group). Using the mean dwell time as a conversion factor,
the absolute risk increase was 0.77 per 2732 PVCs (10 000 PVC
days divided by 3.66 days) and the inverse yields the number
needed to harm, which was approximately 3500 PVCs.
(Reprinted) JAMA Internal Medicine April 2022 Volume 182, Number 4 455
 Letters
Conversely, this means that to prevent 1 case of BSI, 3500 PVCs
would need to be routinely replaced. In addition to patient dis-
comfort, routine replacement increases consumption of re-
sources (eg, device cost, employee time), yet surprisingly, does
not affect thrombophlebitis rate.2,3 Taken together, the deci-
sion to adopt a routine rather than clinically indicated PVC re-
placement policy should be considered in the context of the
high number needed to harm.
Second, it appears that the intervention effect was driven
by the increase in patients with longer dwell time ( >4 days),
from 10.9% to 20.4%.1 The authors’ conclusion is supported
by data in eTable 4 in the Supplement, which show that long
dwell times did not differ significantly between the baseline
and intervention group. However, the distribution of inser-
tion site changed substantially (elbow and wrist instead of fore-
arm) between the baseline and intervention periods. Further-
more, the distribution did not revert when the original policy
was reinstated, suggesting that the insertion site used was in-
dependent from the “permissive” policy and raising the pos-
sibility of latent variables.
Lastly, the clinically indicated replacement group was not
defined by whether the patients had certain indications but
rather by the hospital policy at the time. It would be of inter-
est to report the specific indications that prompted PVC re-
placement (or lack of), especially compared with the entire co-
hort. Did the presence of certain clinical indications, such as
thrombophlebitis or fluid infiltration, increase the risk for BSI?
Tomer Meirson, MD, PhD
Adam Goldman, MD, MPH
David Bomze, MD, MPH
Author Affiliations: Davidoff Cancer Center, Rabin Medical Center−Beilinson
Hospital, Petah Tikva, Israel (Meirson); Sackler Faculty of Medicine, Tel Aviv
University, Tel Aviv, Israel (Goldman, Bomze).
Corresponding Author: David Bomze, MD, MPH, Sackler Faculty of Medicine,
Tel Aviv University, Klatzkin 45 St, Tel Aviv 6997801, Israel (davidbomze@
gmail.com).
Published Online: February 14, 2022. doi:10.1001/jamainternmed.2021.8301
Conflict of Interest Disclosures: None reported.
1. Buetti N, Abbas M, Pittet D, et al. Comparison of routine replacement with
clinically indicated replacement of peripheral intravenous catheters. JAMA
Intern Med. 2021;181(11):1471-1478. doi:10.1001/jamainternmed.2021.5345
2. Webster J, Osborne S, Rickard CM, Marsh N. Clinically-indicated replacement
versus routine replacement of peripheral venous catheters. Cochrane Database
Syst Rev. 2019;1(1):CD007798. doi:10.1002/14651858.CD007798.pub5
3. Vendramim P, Avelar AFM, Rickard CM, Pedreira MDLG. The RESPECT
trial—replacement of peripheral intravenous catheters according to clinical
reasons or every 96 hours: a randomized, controlled, non-inferiority trial. Int J
Nurs Stud. 2020;107:103504. doi:10.1016/j.ijnurstu.2019.103504
To the Editor Dr Buetti and colleagues provided reassuringly low
rates of bloodstream infection (BSI) of less than 1 case per
10 000 peripheral intravenous catheter (PVC) regardless of re-
moval policy.1 However, routine replacement was associated
with statistically fewer PVC-BSI cases compared with clini-
cally indicated replacement (0.005% [46 of 130 779 PVCs] vs
0.035% [15 of 281 852]). Their finding may be subject to a type
1 error and is the inverse of a meta-analysis of 9 well de-
signed, randomized clinical trials (RCTs) that found slightly
456 JAMA Internal Medicine April 2022 Volume 182, Number 4 (Reprinted)
© 2022 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a Universidad de Concepcion (UDEC) User on 11/28/2022
fewer PVC-BSI cases with clinically indicated replacement than
with routine replacement (0.028% [1 of 3590 patients] vs
0.053% [2 of 3733 patients]).2,3
In the study by Dr Buetti and colleagues1 and in the RCTs,2,3
the difference in BSI cases with routine vs clinically indicated
PVC removal was consistently very small, ruling out a larger
effect size of the intervention. Two-thirds of PVC-BSIs oc-
curred on days 1 to 5, without a linear or exponential increase
in per-day risk during the catheter dwell.
The diagnostic definition used by Dr Buetti and colleagues1
was a composite of catheter-related BSI (requiring microbio-
logic confirmation of the PVC as the source) and catheter-
associated BSI (more subjective surveillance-based) defini-
tions—a breakdown would have allowed comparison with the
RCTs, which used catheter-related BSI.4 Furthermore, it would
be valuable to know if assessors were blinded and if inter-
rater reliability was assessed.
No information was provided by the authors on how the
policy change was implemented.1 For staff unaccustomed to
assessing what is or is not a clinically indicated removal of a
PVC, a supportive and structured transition is necessary.5 Were
nursing and medical staff educated to guide appropriate re-
moval decisions and were they empowered to initiate re-
moval? Was dressing durability ensured?
In well conducted RCTs, measured and unmeasured con-
founders are equally distributed between arms, assuring read-
ers of the overall study findings. In this study,1 we did not know
the effects of important factors such as cancer diagnosis, im-
munosuppression, or difficult PVC insertion. We caution
against using observational studies to inform and/or change
practice, particularly when a reduced risk of infection of 1 (at
most) per 10 000 PVC days would incur substantial eco-
nomic, staff time, and patient experience costs.
Claire M. Rickard, RN, PhD
David L. Paterson, MBBS, PhD
Vineet Chopra, MD, MSc
Author Affiliations: Herston Infectious Diseases Institute, Metro North Health,
School of Nursing, Midwifery, and Social Work, University of Queensland Centre
for Clinical Research, Herston, Queensland, Australia (Rickard, Paterson); School
of Medicine, University of Colorado, Aurora (Chopra).
Corresponding Author: Claire M. Rickard, RN, PhD, Herston Infectious Diseases
Institute, Metro North Health, School of Nursing, Midwifery, and Social Work,
University of Queensland Centre for Clinical Research, Room 306, Herston, QLD
4006, Australia (c.rickard@uq.edu.au).
Published Online: February 14, 2022. doi:10.1001/jamainternmed.2021.8304
Conflict of Interest Disclosures: Dr Rickard reported grants and consulting
payments to her institution for speaking engagements from 3M, Becton
Dickinson, Cardinal Health, and Eloquest, outside the submitted work.
Dr Paterson reported funding from AstraZeneca, Leo Pharmaceuticals, Bayer,
GlaxoSmithKline, Cubist, Venatorx, and Accelerate; board membership for
Janssen, Entasis, Qpex, Merck, Shionogi, Achaogen, AstraZeneca, Leo
Pharmaceuticals, Bayer, GlaxoSmithKline, Cubist, Venatorx, and Accelerate;
grants from Pfizer, Shionogi, and Merck; speaking fees from Pfizer, all outside
the submitted work. No other disclosures were reported.
1. Buetti N, Abbas M, Pittet D, et al. Comparison of routine replacement with
clinically indicated replacement of peripheral intravenous catheters. JAMA
Intern Med. 2021;181(11):1471-1478. doi:10.1001/jamainternmed.
2021.5345
2. Rickard CM, Webster J, Wallis MC, et al. Routine versus clinically indicated
replacement of peripheral intravenous catheters: a randomised controlled
jamainternalmedicine.com

equivalence trial. Lancet. 2012;380(9847):1066-1074. doi:10.1016/S0140-6736
(12)61082-4
3. Webster J, Osborne S, Rickard CM, Marsh N. Clinically-indicated replacement
versus routine replacement of peripheral venous catheters. Cochrane Database
Syst Rev. 2019;1:CD007798. doi:10.1002/14651858.CD007798.pub5
4. Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the
diagnosis and management of intravascular catheter-related infection. Clin
Infect Dis. 2009;49(1):1-45. doi:10.1086/599376
5. Takashima M, Cooke M, DeVries M, et al. An implementation framework for
the clinically indicated removal policy for peripheral intravenous catheters.
J Nurs Care Qual. 2021;36(2):117-124. doi:10.1097/NCQ.0000000000000507
To the Editor We congratulate Buetti and colleagues1 on their
excellent analysis of whether routine replacement (RR) of pe-
ripheral intravenous catheters (PVCs) is superior to clinically
indicated replacement. The primary end point of the study was
PVC-related bloodstream infections (PVC-BSIs), which we agree
is a better measure than the phlebitis rates used by previous
randomized clinical trials. Phlebitis is usually associated with
mechanical (PVC insertion) or chemical (administered drugs)
irritation, rather than infectious causes. Changing recommen-
dations based on infectious complications is relevant be-
cause they are burdensome for patients as well as for health
care systems.
It is necessary to translate the study results into a clini-
cally understandable measure. The number needed to treat is
an epidemiologic measure indicating the number of people re-
quired to undergo an intervention to prevent an adverse ef-
fect in 1 individual.2 It is defined as the inverse of absolute risk
reduction, which is the difference between incidence in the
control (clinically indicated replacement) vs the experimen-
tal group (RR). The data from this study suggest that 3279 RR-
PVCs would have to be performed to prevent a single case of
PVC-BSI. This indicates that the number needed to treat for
PVC-BSIs using RR does not warrant a policy change, espe-
cially given its disadvantages. The implementation of RR is as-
sociated with higher health care costs and increased staff
workload.3 Additionally, with RR, exposure to unnecessary in-
vasive procedures is increased, negatively affecting patient
safety and increasing their discomfort. Also, RR may threaten
health care staff safety by increasing the likelihood of needle-
stick injuries.
It is also important to consider the context of the patients
for whom PVC-BSIs events occur because context may help
identify prevention methods. For example, PVCs that are in-
serted in emergency settings are known to have a higher risk
of PVC-BSIs because of the difficulty of maintaining asepsis;4
thus, RR-PVC is recommended within 24 hours in these set-
tings. Moreover, there are simple measures that can be used
to lower infection rates and decrease patient and health sys-
tem burden; for example, regular PVC site examination and im-
mediate removal of PVC when clinical signs and/or symp-
toms are suggestive of infection.
Buetti and colleagues1 have produced a commendable
analysis that argues for changing policy to routinely replac-
ing PVCs. Nevertheless, it is important to consider whether a
change would provide real improvement in patient care. We
must take into account the entire picture, including the num-
ber needed to treat for PVC-BSIs using RR, the disadvantages,
jamainternalmedicine.com
© 2022 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a Universidad de Concepcion (UDEC) User on 11/28/2022
Letters
and the possibility of implementing simple measures to pre-
vent serious PVC-BSIs.
Agastya Patel, MD
Piotr Spychalski, MD, PhD
Jarek Kobiela, MD, PhD, MBA
Author Affiliations: Department of General, Endocrine, and Transplant Surgery,
Medical University of Gdańsk, Gdańsk, Poland.
Corresponding Author: Agastya Patel, MD, Department of General, Endocrine,
and Transplant Surgery, Medical University of Gdańsk, ul. Smoluchowskiego 17,
80-214 Gdańsk, Poland (agastyap24@gumed.edu.pl).
Published Online: February 14, 2022. doi:10.1001/jamainternmed.2021.8307
Conflict of Interest Disclosures: None reported.
1. Buetti N, Abbas M, Pittet D, et al. Comparison of routine replacement with
clinically indicated replacement of peripheral intravenous catheters. JAMA
Intern Med. 2021;181(11):1471-1478. doi:10.1001/jamainternmed.2021.5345
2. Citrome L, Ketter TA. When does a difference make a difference?
interpretation of number needed to treat, number needed to harm, and
likelihood to be helped or harmed. Int J Clin Pract. 2013;67(5):407-411.
doi:10.1111/ijcp.12142
3. Olivier RC, Wickman M, Skinner C, Ablir L. The impact of replacing peripheral
intravenous catheters when clinically indicated on infection rate, nurse
satisfaction, and costs in CCU, step-down, and oncology units. Am J Infect Control.
2021;49(3):327-332. doi:10.1016/j.ajic.2020.07.036
4. Mermel LA. Short-term peripheral venous catheter-related bloodstream
infections: a systematic review. Clin Infect Dis. 2017;65(10):1757-1762.
doi:10.1093/cid/cix562
In Reply We read with great interest the correspondence from
Dr Meirson and colleagues, Dr Patel and colleagues, and
Dr Rickard and colleagues in response to our original investi-
gation on replacement of peripheral intravenous catheters
(PVCs).1 Dr Meirson and colleagues correctly point out that the
number needed to harm was very high. We discussed this point
in the article and whether an incidence increase of 1 PVC-
associated bloodstream infection (BSI) per 10 000 days of PVC
use (or of 0.77 PVC-BSI per 2732 PVCs placed) justifies rou-
tine replacement. From an ecologic perspective, approxi-
mately 1.2 billion PVCs are placed or replaced annually; thus,
even though the outcome is rare, a large number of patients
contract a PVC-BSI per year in the world. Moreover, during the
intervention period we detected a high proportion of Staphy-
lococcus aureus PVC-BSIs, which is associated with high mor-
tality. Interestingly, a recent study reported that infections
caused by S aureus and gram-negative bacilli were more fre-
quently detected in PVC-BSIs compared with central venous
catheter-related BSIs.2
In summary, the following arguments are against routine
replacement: (1) the low incidence of PVC-BSI and high num-
ber needed to harm; (2) increased cost of material and work-
ing time; (3) patient discomfort, particularly for those with de-
creased venous capital; (4) increased risk of needlestick injuries
among health care workers; and (5) the lack of effectiveness
for phlebitis prevention, the risk of which appears to be simi-
lar among both routine PVC replacement and clinically indi-
cated PVC replacement groups.3 On the other hand, the fol-
lowing arguments favor routine replacement: (1) the total
burden of PVC-BSI given the high number of PVCs placed and
replaced; (2) a nonnegligible proportion of PVC-BSIs caused
(Reprinted) JAMA Internal Medicine April 2022 Volume 182, Number 4 457
 Letters
by S aureus1 and other virulent microorganisms, particularly
among the clinically indicated group; and (3) prevention of
other complications, such as fluid infiltration into surround-
ing tissues and blockage-related catheter failure.3 Only 1 trial
reported mortality and found no clear difference in the inci-
dence of mortality when clinically indicated replacement was
compared with routine replacement.4
In light of these considerations, we cannot conclude that
PVCs must be routinely replaced in every health care facility;
however, the cutoff of clinical relevance remains to be eluci-
dated, as do other considerations (eg, cost effectiveness). Our
large study1 allowed for a reality check that challenges the con-
clusions of small randomized clinical trials.
Dr Patel and colleagues highlight simple actions that can
be used to reduce infection rates, eg, that routine site exami-
nation and immediate removal of PVC when signs or symp-
toms are clinically suggestive of infection—presuming that exit-
site surveillance is sufficient for monitoring serious infection.
Although surveillance is essential to detecting exit-site infec-
tions, local symptoms may not be reliably indicative of
PVC-BSI.5 We agree that site examination should be per-
formed routinely. However, although we could not measure
this formally, there was probably heterogeneity in the fre-
quency and intensity of exit-site examination in the study co-
hort, and the shift toward clinically indicated replacement may
have adversely affected the health care workers’ attention to
the PVCs.1
Dr Rickard and colleagues correctly point out that the com-
parison of the study results1 with those of other randomized
clinical trials should be interpreted with caution because we
used a composite outcome (ie, catheter-related and catheter-
associated BSI). Also, assessors were not blinded. However, sur-
veillance was part of routine prospective BSI surveillance,
which has been in place since 1998. During the clinically in-
dicated and routine replacement period, only 4 and 3 BSIs, re-
spectively, were formally related to PVCs; tip cultures of pe-
ripheral lines were not routinely performed. The policy change
was introduced by the vascular access team via the nursing hi-
erarchy, and the data showed that the new protocol was taken
up rapidly; however, it was not accompanied by any formal in-
stitutional training on the appropriate timing of catheter re-
moval. Training could have curtailed the substantial increase
in PVC-associated BSI.
In conclusion, the jury is still out on whether PVCs should
be replaced routinely or only when clinically indicated. While
further evidence is being collected and analyzed, policy mak-
ers must weigh the harms and benefits of the options when
implementing their local strategies.
Niccolò Buetti, MD, MSc, PhD
Mohamed Abbas, MD, MSc
Walter Zingg, MD
Author Affiliations: Infection Control Program and World Health Organization
Collaborating Center on Patient Safety, University of Geneva Hospitals and
Faculty of Medicine, Geneva, Switzerland (Buetti, Abbas, Zingg); Infection,
Antimicrobials, Modelling, Evolution (IAME) Laboratory, University of Paris,
Paris, France (Buetti); Division of Infectious Diseases and Hospital
Epidemiology, University Hospital Zurich, Zurich, Switzerland (Zingg).
458 JAMA Internal Medicine April 2022 Volume 182, Number 4 (Reprinted)
© 2022 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a Universidad de Concepcion (UDEC) User on 11/28/2022
Corresponding Author: Niccolò Buetti, MD, MSc, PhD, Infection Control
Program and World Health Organization Collaborating Center on Patient Safety,
University of Geneva Hospitals and Faculty of Medicine, Rue
Gabrielle-Perret-Gentil 4, 1205 Geneva, Switzerland (niccolo.buetti@hcuge.ch).
Published Online: February 14, 2022. doi:10.1001/jamainternmed.2021.8310
Conflict of Interest Disclosures: Dr Buetti reported a fellowship grant from the
Swiss National Science Foundation (No. P4P4PM19 4449). No other disclosures
were reported.
1. Buetti N, Abbas M, Pittet D, et al. Comparison of routine replacement with
clinically indicated replacement of peripheral intravenous catheters. JAMA
Intern Med. 2021;181(11):1471-1478. doi:10.1001/jamainternmed.2021.5345
2. Tatsuno K, Ikeda M, Wakabayashi Y, Yanagimoto S, Okugawa S, Moriya K.
Clinical features of bloodstream infections associated with peripheral versus
central venous catheters. Infect Dis Ther. 2019;8(3):343-352. doi:10.1007/
s40121-019-00257-6
3. Webster J, Osborne S, Rickard CM, Marsh N. Clinically-indicated replacement
versus routine replacement of peripheral venous catheters. Cochrane Database
Syst Rev. 2019;1:CD007798. doi:10.1002/14651858.CD007798.pub5
4. Rickard CM, Webster J, Wallis MC, et al. Routine versus clinically indicated
replacement of peripheral intravenous catheters: a randomised controlled
equivalence trial. Lancet. 2012;380(9847):1066-1074. doi:10.1016/S0140-6736
(12)61082-4
5. Buetti N, Ruckly S, Lucet JC, et al. Local signs at insertion site and
catheter-related bloodstream infections: an observational post hoc analysis
using individual data of four RCTs. Crit Care. 2020;24(1):694. doi:10.1186/
s13054-020-03425-0
Strengthening a Study of Diabetes Progression
After Statin Use
To the Editor We were interested by the recent article from Dr
Mansi and colleagues who reported a significant progression
of type 2 diabetes after initiation of statin therapy compared
with initiation of proton pump inhibitor or H2-blocker.1 In their
discussion, the authors acknowledged that the composite out-
come for diabetes progression (ie, therapy escalation, persis-
tent hyperglycemia >200 mg/dL, etc) was a study limitation;
indeed data regarding hemoglobin A1C (HbA1c) levels were not
collected. Although research shows that HbA 1c levels in-
crease marginally after initiating a statin (+0.1%),2 the differ-
ence may not be negligible when drug exposure is prolonged
for years in patients with diabetes of long duration.
The report from Dr Mansi and colleagues1 prompted us to
test whether HbA1c levels and the duration of diabetes dif-
fered for the patients we have treated by a statin. Among 905
patients with type 2 diabetes (mean [SD] age, 62 [10] years;
57.3% men), we treated 575 (63.5%) with statin therapy. The
mean (SD) HbA1c levels among them were: 8.9% (2.0%) vs 8.8%
(1.8%) without statin use (P = .42). The main expression of a
higher glucose exposure among the patients receiving statin
therapy was a longer mean (SD) duration of diabetes: 15 (9)
years vs 11 (10) years (P < .001).
Diabetes duration was not reported in the article,1 and du-
ration was not included in the propensity scores to match statin
users with comparators. The matching for microvascular dia-
betic complications was not comprehensive, eg, only 3.3% of
participants had ophthalmic manifestations. We assume that
many diabetic retinopathies were missed.
We suggest that higher LDL cholesterol levels (111 vs
96 mg/dL) may not have been the sole reason for statin
initiation among the study cohort. 1 If these patients had
long-standing diabetes, it may have led to (1) lower LDL
jamainternalmedicine.com