This study assessed using intravenous sodium bicarbonate to verify proper intravenous catheter placement in mechanically ventilated children. The researchers injected diluted sodium bicarbonate or saline through catheters in 18 anesthetized children and measured end-tidal carbon dioxide levels. Injecting 2.1% sodium bicarbonate significantly increased end-tidal carbon dioxide within 3 breaths, correctly identifying catheter placement. A less concentrated solution caused a smaller increase. Saline did not affect carbon dioxide levels. Blood tests found no clinically significant changes. The study suggests intravenous sodium bicarbonate injection can safely confirm pediatric intravenous catheter function by detecting increased exhaled carbon dioxide.

1. Society for Pediatric Anesthesia
Section Editor: Peter J. Davis
Intravenous Sodium Bicarbonate Verifies Intravenous
Position of Catheters in Ventilated Children
Ilan Keidan, MD,*† Erez Ben-Menachem, MBCHB, FANZCA,* Sno Ellen White, MD,†
and Haim Berkenstadt, MD*
BACKGROUND: Vascular access in children carries a significant risk of accidental extravasation
of IV fluids and medications with the potential for tissue injury. In this prospective controlled
study we assessed the diagnostic utility of using IV diluted sodium bicarbonate to confirm
placement of IV catheters in ventilated children. Diluted sodium bicarbonate was created using
undiluted standard 8.4% (1 mEq/mL) sodium bicarbonate mixed in a 1:3 and 1:5 ratio with
sterile water to achieve a final diluted concentration of 2.1% (0. 25 mEq/mL) and 1.05% (0.125
mEq/mL) sodium bicarbonate, respectively.
METHODS: In 18 ASA I–II mechanically ventilated children ages 1 to 8 years, the effects of 1 mL/kg
of dilute 2.1%, 1.05% sodium bicarbonate, or 0.9% normal saline, injected in a randomized order,
were analyzed. All children had oxygen saturation, arterial blood pressure, electrocardiograph, and
end-tidal carbon dioxide (ETCO2) monitoring. In addition, venous blood samples were taken before
injection and 10 minutes after the final injection for analysis of venous blood pH and electrolytes.
RESULTS: In children, IV diluted 2.1% sodium bicarbonate resulted in significantly increased ETCO2
(mean of 32.8 Ϯ 3.4 mm Hg to 39.0 Ϯ 3.5 mm Hg, P Ͻ 0.001), a mean increase of 6.2 mm Hg
(95% prediction interval: 4.3 to 8.1 mm Hg) within 3 breaths. Intravenous diluted 1.05% sodium
bicarbonate caused a less pronounced but still significant increase in ETCO2 (33.4 Ϯ 3.8 mm Hg to
36.3 Ϯ 3.4 mm Hg, P Ͻ 0.001), a mean increase of 2.9 mm Hg (95% prediction interval: 1.8 to 4.1
mm Hg) within 3 breaths. Normal saline did not result in any significant changes, with a mean
increase of 0.06 mm Hg (95% prediction interval: Ϫ1.3 to 1.4 mm Hg). Both concentrations of
sodium bicarbonate were easily differentiated from normal saline injection by blinded anesthesiolo-
gists observing the change in ETCO2 values immediately after injection. Analysis of pre- and
postinjection venous pH, bicarbonate, and sodium levels could not detect clinically significant
changes. A small but statistically significant increase in venous bicarbonate was noted.
CONCLUSION: The injection of 2.1% sodium bicarbonate in mechanically ventilated ASA I–II
children identified intravascular placement and patency of an IV catheter by an increase in the
exhaled CO2 concentration. The injections did not have any clinically significant effects on blood
pH, bicarbonate, or sodium concentration. (Anesth Analg 2012;115:909–12)
It is a common clinical problem to be presented with a
pediatric patient with IV access for which there is doubt
about the usability of the catheter. Vascular access is often
bandaged, obscuring clinical assessment, children may not be
capable of verbally communicating pain at injection sites, and
a “twiddler’s syndrome” has been described in which the
child manipulates the catheter, causing it to migrate out of the
vessel.1
Additionally, fluid leakage into surrounding tissue
may initially go unnoticed owing to the distensibility of
subcutaneous tissues in the very young. Even with infusion
devices programmed to alarm in response to increasing
infusion pressures,2
hopefully allowing early identification of
infiltration, the high compliance of subcutaneous tissues in
children leads to failure of the alarm and reduces the effec-
tiveness of this safety feature.3
Moreover, manual IV boluses
may generate higher pressures, resulting in rupture of fragile
vessels. Traditionally, correct IV placement has been con-
firmed by a return flow of blood and gentle flushing of the
cannula with saline and observation of the surrounding
tissues. However, these methods are not consistently accurate
because small-caliber vessels and potentially dehydrated pa-
tients do not allow return flow of blood, and highly compliant
or edematous tissues may obscure extravasation of saline and
give a false sense of security.
Infiltration, the unintentional leakage of IV fluids into
the surrounding tissue, is the most common complication
of vascular access. Extravasation, the leakage of IV medi-
cations into tissues can also occur. Both can cause injury,
and the terms will be used interchangeably in this manuscript.
The incidence of peripheral IV infiltration was reported to be
23%–78% for neonates4
and 11%–28% for children,5,6
whereas
percutaneous central venous catheters were reported to have
an infiltration rate of 1% to 16%.7
Extravasation of IV medi-
cations may result in tissue damage with the potential for skin
From the *Department of Anesthesia and Intensive Care, Sheba Medical
Center, affiliated with the Sackler Faculty of Medicine, Tel Aviv University,
Israel; †Department of Anesthesiology, University of Florida, Gainesville,
Florida.
Accepted for publication March 27, 2012.
Funding: Department funds were used to support this study, specifically for
purchase of drugs and statistical analysis of data.
The authors declare no conflict of interest.
Reprints will not be available from the authors.
Address correspondence to Erez Ben-Menachem, MBCHB, FANZCA, De-
partment of Anesthesia and Intensive Care, Sheba Medical Center, affiliated
with the Sackler Faculty of Medicine, Tel Hashomer 52621, Israel. Address
e-mail to erezben@yahoo.com.
Copyright © 2012 International Anesthesia Research Society
DOI: 10.1213/ANE.0b013e318258023b
October 2012 • Volume 115 • Number 4 www.anesthesia-analgesia.org 909

2. necrosis, with tissue sloughing being reported in as many as
44% of infants with infiltrations.8
Another important consequence of inappropriate place-
ment of IV access is repeated cannulation. This procedure is
stressful for both children and parents and has been rated
as the most distressing part of hospital admissions and the
cause of the most pain while in hospital.9
Against this
backdrop, there are multiple clinical scenarios in which
there is a strong desire to confirm correct functioning of a
doubtful “in situ” catheter.
Given the regular occurrence of the clinical dilemma
already outlined, shortcomings of traditional methods, and
the risk of tissue damage from extravasation, we decided to
test a technique that may differentiate between an infil-
trated IV catheter and a correctly sited catheter.
In this prospective controlled trial, the diagnostic utility
of using IV-injected dilute 2.1% (0. 25 mEq/mL) and 1.05%
(0.125 mEq/mL) sodium bicarbonate in mechanically ven-
tilated children was tested. The technique is based on the
observed phenomenon of increased end-tidal carbon diox-
ide (ETCO2) concentration after the IV administration of
sodium bicarbonate,10
and mechanically ventilated chil-
dren under general anesthesia were chosen because of the
controlled setting and ability to accurately measure ETCO2
and other physiological variables. In an effort to reduce the
sequelae of extravasated sodium bicarbonate, which is
hyperosmolar and highly alkaline, it was diluted with
sterile water, and its physicochemical properties were
tested. We have already investigated the value of the
technique in mechanically ventilated adult patients and
performed a safety study of subcutaneously injected so-
dium bicarbonate in rats; the results have recently been
published.11
Our aim in this study was to identify whether
a lower concentration of sodium bicarbonate would still
have diagnostic utility, investigate the effects of a bolus of
diluted sodium bicarbonate on the unique physiology of
children, and measure possible serum electrolyte changes.
METHODS
Institutional ethics committee approval was granted for
this study, and prior written consent was obtained from the
legal guardians of all children before enrollment. Eighteen
ASA I–II patients ages 1 to 8 years undergoing elective
surgery under general anesthesia with tracheal intubation
and controlled mechanical ventilation were studied. Gen-
eral anesthesia was induced using a sevoflurane inhaled
technique. Midazolam premedication was given at the
discretion of the treating anesthesiologist. After induction
of general anesthesia, a free-flowing 22-gauge IV cannula
(BD Venflon, Helsingborg, Sweden) was placed in the
upper limb of the patient. Correct IV placement was
confirmed by free aspiration of blood. Rocuronium 0.5
mg/kg was then given, and the trachea was intubated with
an appropriately sized cuffed endotracheal tube. The cuff
was inflated until there was no detectable leak, with cuff
pressure kept below 20 cmH2O. The ventilator was set at
tidal volumes of 10 mL/kg of body weight, positive end-
expiratory pressure of 5 cm/H2O, and respiratory rate (RR)
of 15 breaths per minute. After tracheal intubation, and
before surgery, a period of 5 minutes was allowed, at the
end of which baseline ETCO2 was established. The patient
then received a 1 mL/kg bolus of 0.9% normal saline, 2.1%
sodium bicarbonate, or 1.05% sodium bicarbonate. Sodium
bicarbonate was diluted with sterile water and was ana-
lyzed for pH (Glass Electrode Radiometer, Copenhagen,
Denmark) and osmolarity (Osmostation, OM 6050 A. Men-
arini Industries, Florence, Italy). Each patient received all 3
solutions given as a bolus and injected in a randomized
order in 5-minute intervals. The injections were given
through the cannula side-port close to the insertion site.
Injection pressures were not measured, and no flush fluids
were given. All injections and data collection occurred after
induction of anesthesia and hemodynamic and ventilatory
stabilization and before commencement of surgery. Injec-
tions were performed by an anesthesiologist not involved
in the patient’s anesthetic care so that the treating anesthesi-
ologist remained blinded to the injectate. The data collected
included patient demographics, ETCO2 values during 11
consecutive breaths after the injection (S/5 anesthesia moni-
tor; Datex Ohmeda now a division of GE Healthcare,
Waukesha, WI), heart rate for 5 minutes after injection, and
noninvasive arterial blood pressure measured 2.5 and 5
minutes after the injection. An increase of 10% in the
ETCO2 from baseline was considered a positive result.
Venous blood samples were taken at baseline and 10
minutes after all 3 solutions had been injected. These
venous blood samples were analyzed for pH and electro-
lyte levels. The blinded treating anesthesiologist was asked
after each injection to determine whether sodium bicarbonate
or normal saline had been given. A power analysis was
conducted, with ␣ ϭ 0.05, 1 Ϫ ␤ ϭ 0.9, difference to be
detected ϭ 2 mm Hg, with an SD of differences of 1 mm Hg.
The minimum required sample size for a 2-sided paired t
test ϭ 5 pairs, with 18 pairs allowing the power of the test to
be 100%. Paired comparisons between ETCO2 before and after
injection as well as blood pH and electrolytes were made
using the Student paired t test. A ␹-square test of goodness of
fit was performed to determine whether the blinded observers
successfully detected the sodium bicarbonate or normal saline
injections in the 18 patients. Data were analyzed using SPSS
software (version 19; IBM Corp., Somers, NY).
RESULTS
The physical characteristics of the injected diluted sodium
bicarbonate are outlined in Table 1. Diluting sodium bicar-
bonate with sterile water decreased the osmolarity of the
injectate from 2000 mOsm/L for the undiluted sodium
bicarbonate to 447 mOsm/L for the diluted 2.1% sodium
bicarbonate and 250 mOsm/L once diluted to 1.05%. The
dilution did not have any effect on the pH of the injectate.
Eighteen patients ages 1 to 8 years old were enrolled in
this study (mean age: 5 Ϯ 2 years). Mean patient weight
was 22 Ϯ 8 kg (range 9 to 34 kg). Peak ETCO2 increased by
Table 1. The Physicochemical Characteristics of
Diluted Sodium Bicarbonate
Sodium bicarbonate
(diluted with sterile water)
Undiluted
(8.4%)
1:1
(4.2%)
1:3
(2.1%)
1:5
(1.05%)
Osmolarity (mOsm/L) 2000 818 447 250
pH 8.6 8.68 8.72 8.77
A Technique to Confirm IV Line Position in Children
910 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

3. Ͼ10% in all patients after injection with dilute 2.1% sodium
bicarbonate (from a mean value of 32.7 Ϯ 3.4 mm Hg to
39.0 Ϯ 3.5 mm Hg, P Ͻ 0.001). The peak ETCO2 occurred in
the second breath after the completion of the injection in 16
out of 18 patients and in the third breath in the remaining
2 of 18 patients. Injection with diluted 1.05% sodium
bicarbonate resulted in a significant increase in ETCO2
(from a mean value of 33.4 Ϯ 3.8 mm Hg to 36.3 Ϯ 3.4 mm
Hg, P Ͻ 0.01), but reached a 10% increase in only 5 of the
18 patients (28%). Normal saline did not result in a signifi-
cant increase in ETCO2 levels (Fig. 1). The blinded anesthe-
siologist could consistently detect the injection with 2.1%
sodium bicarbonate, but the detection rate with 1.05%
bicarbonate was 80%. All other physiological variables (RR,
heart rate, blood pressure) remained unchanged. After
injection, venous blood samples showed no increase in
blood pH, serum sodium, or ionized calcium, and a small
statistically significant, but clinically insignificant, increase
in bicarbonate (Table 2).
DISCUSSION
This study describes the novel use of a clinically observed
phenomenon to confirm the correct placement and viability
of an IV catheter placed in mechanically ventilated, endo-
tracheally intubated pediatric patients. ETCO2 increase
after injection of 2.1% sodium bicarbonate appears to
reliably confirm IV placement of an IV catheter. The use of
1.05% sodium bicarbonate is less reliable. Care should be
taken if testing an IV cannula previously or concurrently
being used for medications, because sodium bicarbonate
may be incompatible with certain drugs, including epi-
nephrine, norepinephrine, dobutamine, midazolam, cal-
cium salts, and magnesium salts.
The complications of extravasation from a nonfunction-
ing IV catheter can be serious. The American Society of
Anesthesiologists Closed Claims Project database reflects
this, with 2.1% of all claims (all ages) between 1975 and
2000 being related to peripheral catheters; of these, 28%
were due to skin slough or necrosis and 17% were due to
swelling, inflammation, or infection.12
It was calculated
that 55% of all complications were the result of extravasa-
tion of drugs or fluid. Given the seriousness of the problem
of inadvertent extravasation with the potential for injury,
increased costs, and risk of litigation, it would seem sen-
sible to develop techniques, such as outlined above, that
assist in preventing such negative outcomes. All IV medi-
cations and fluids can cause tissue damage, but certain
substances are associated with a greater risk of tissue necro-
sis.13
Moreover, though this study investigated confirmation
of placement and patency of peripheral IV cannulas, the
technique may be equally useful for central venous access
catheters, and further studies in this setting are warranted.
The consequences of misplaced or migrated central vascular
access can be catastrophic with the potential for infusion of
fluids into the pleural or abdominal cavity. Cardiovascular
collapse could potentially occur should inotropic infusions be
given through a catheter not positioned intravascularly. Over-
all, given the frequency of the problem, the potential compli-
cations and associated costs, and the distress caused to
patients and families, a simple clinical test such as described
in this study could beneficially impact health care spending
attributable to hospital-acquired conditions and improve con-
sumer satisfaction assessments.
Figure 1. Bland–Altman plot of increase in
end-tidal CO2 against baseline, pre-, and postin-
jection (2.1% sodium bicarbonate, 1.05% so-
dium bicarbonate, 0.9% normal saline).
Table 2. Venous Blood Samples Before and After
the Injection of the Study Injectates
Before
injection
After
injection Significance
Blood pH 7.39 Ϯ 0.04 7.39 Ϯ 0.05 NS
Sodium bicarbonate
(mEq/L)
20.75 Ϯ 1.4 22.19 Ϯ 1.8 Ͻ0.05
Sodium (mEq/L) 135.5 Ϯ 2.1 135.6 Ϯ 1.7 NS
Ionized calcium (mmol/L) 1.1 Ϯ 0.4 1.05 Ϯ 0.5 NS
NS ϭ not significant.
October 2012 • Volume 115 • Number 4 www.anesthesia-analgesia.org 911

4. Any solution with osmolarity greater than plasma osmo-
larity causes damage by exerting an osmotic pressure leading
to a compartment syndrome. Similarly highly alkaline solu-
tions with pH above 9.0 can cause direct cellular injury, and it
is recommended that infused solutions should have a pH
between 5 and 9.14
Potent vasoconstictors can also cause
necrosis secondary to tissue ischemia. Drugs commonly used
in the perioperative setting with the potential to cause tissue
damage are listed in Table 3, although other drugs not
commonly associated with tissue damage still have the po-
tential to cause injury, including propofol.15
Vesicants are substances that have physical or chemical
properties that make them capable of causing inflamma-
tion, pain, and blistering of tissues, potentially causing
tissue death and necrosis, with injury to surrounding tissue
if wrongly administered subcutaneously. Sodium bicarbon-
ate is considered a vesicant secondary to its high pH and
high osmorality. Although we did not specifically investi-
gate the effect of subcutaneous injection of dilute sodium
bicarbonate in pediatric patients, its minimally useful con-
centration is likely to enhance the safety of this technique
by decreasing the osmolarity of the injectate. The safety of
subcutaneous injection of dilute 4.2% sodium bicarbonate
has been tested in a rat model and appears safe.11
Notably,
a useful minimum concentration (2.1% sodium bicarbon-
ate) has been found for this technique in pediatric patients.
Conspicuously, the peak increase in ETCO2 was seen at the
second to third breath, in contrast to the seventh breath in
adults.11
It is likely that this earlier increase is the result of the
higher relative cardiac output and faster circulation time seen
in children. It is likely that the technique is equally applicable
to children whose lungs are ventilated via a laryngeal mask;
however, spontaneously breathing patients may not exhibit
the same CO2 increase because they may partially compensate
with an increased RR. Further studies that observe RR and
exhaled CO2 in pediatric patients with lung disease, sponta-
neously breathing, or awake pediatric patients may be war-
ranted. Additional investigation may be useful to confirm the
optimal volume, concentration, and total dose of sodium
bicarbonate for this technique. Reassuringly, the 2.1% sodium
bicarbonate bolus (0.25 mEq/kg), even if repeated several
times a day, is well below the recommended 8 mEq/kg/day
sodium bicarbonate dose limit for children younger than 2
years of age. Until further evidence suggests otherwise, we
would consider contraindications to include premature in-
fants, especially those with intraventricular hemorrhage, and
children with increased intracranial pressure.
In summary, we present a useful technique, using 1
mL/kg of dilute 2.1% sodium bicarbonate and the resultant
ETCO2 increase, to confirm correct intravascular placement
of peripheral IV cannulas in anesthetized, mechnically
ventilated patients. No adverse events or complications
were detected. This simple test may aid in reducing the
complications of infiltration, and the patient and family
distress caused by the common clinical scenario of the
“doubtful” peripheral IV cannula.
DISCLOSURES
Name: Ilan Keidan, MD.
Contribution: This author helped design the study, implement
the study, and compose the manuscript.
Name: Erez Ben-Menachem, MBCHB, FANZCA.
Contribution: This author helped implement the study and
compose the manuscript.
Name: Sno Ellen White, MD.
Contribution: This author helped compose the manuscript and
the revision.
Name: Haim Berkenstadt, MD.
Contribution: This author helped implement the study and
compose the manuscript.
This manuscript was handled by: Peter J. Davis, MD.
REFERENCES
1. Servetar EM. A case report of Twiddler’s syndrome in a
pediatric patient. J Pediatr Oncol Nurs 1992;9:25–8
2. Perdue MB. Intravenous complications. In: Hankins L, Lon-
sway RA, Hendrick C, Perdue MB, eds. Infusion Therapy in
Clinical Practice. 2nd ed. St. Louis, MO: Saunders, 2011:418–45
3. Hee HI, Lim SL, Tan SS. Infusion technology: a cause for alarm.
Pediatr Anaesth 2002;12:780–5
4. Franck LS, Hummel D, Connell K, Quinn D, Montgomery J.
The safety and efficacy of peripheral intravenous catheters in
ill neonates. Neonatal Netw 2001;20:33–8
5. Kassner E. Evaluation and treatment of chemotherapy extrava-
sation injuries. J Pediatr Oncol Nurs 2000;17:135–48
6. Garland JS, Dunne WM Jr, Havens P, Hintermeyer M, Bozzette
MA, Wincek J, Bromberger T, Seavers M. Peripheral intrave-
nous catheter complications in critically ill children: a prospec-
tive study. Pediatrics 1992;89:1145–50
7. Wynsma LA. Negative outcomes of intravascular therapy in
infants and children. AACN Clin Issues 1998;9:49–63
8. Pettit J, Hughes K. Intravenous extravasation: mechanisms,
management, and prevention. J Perinat Neonatal Nurs 1993;
6:69–78
9. Cummings EA, Reid GJ, Finley GA, McGrath PJ, Ritchie JA.
Prevalence and source of pain in pediatric inpatients. Pain
1996;68:25–31
10. Okamoto H, Hoka S, Kawasaki T, Okuyama T, Takahashi S.
Change in end-tidal carbon dioxide tension following sodium
bicarbonate administration: correlation with cardiac output
and haemoglobin concentration. Acta Anaesthesiol Scand
1995;39:79–84
11. Keidan I, Ben-Menachem E, Barzilai A, Nur I, Berkenstadt H.
Intravenous sodium bicarbonate verifies intravenous position of
catheters in ventilated patients. Anesth Analg 2011;113:279–81
12. Bhananker SM, Liau DW, Kooner PK, Posner KL, Caplan RA,
Domino KB. Liability related to peripheral venous and arterial
catherization: a closed claims analysis. Anesth Analg 2009;
109:124–9
13. Doellman D, Haaway L, Bowe-Geddes LA, Franklin M, Le-
Donne J, Papke-O’Donnell L, Pettit J, Schulmeister L, Stranz M.
Infiltration and extravasation: update on prevention and man-
agement. J Infus Nurs 2009;32:203–11
14. Vesely TM, Stranz M, Masoorli S, Hadaway LC. The diverse
and conflicting standards and practices in infusion therapy.
J Vasc Devices 2002;7:9–25
15. Roth W, Eschertzhuber S, Gardetto A, Keller C. Extravasation
of propofol is associated with tissue necrosis in small children.
Pediatr Anesth 2006;16:887–9
Table 3. Drugs Used in Anesthesia with the
Potential to Cause Tissue Damage
Hyperosmolar
agents
Acids/
alkalis
Vasocontrictive
agents
Hypertonic saline Thiopental Epinephrine
Potassium chloride Diazepam Norepinephrine
Calcium chloride Phenytoin Dobutamine
Calcium gluconate Vancomycin Metaraminol
Mannitol Vasopression
Sodium bicarbonate Phenylephrine
Glucose Ͼ10%
A Technique to Confirm IV Line Position in Children
912 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA