Br. J. Anaestk.
(1985), 57, 220-233
INFUSION THROMBOPHLEBITIS
G. B. H. LEWIS AND J. F. HECKER
An i.v. infusion is the most commonly performed
surgical procedure in hospital wards (Eremin and
Marshall, 1977) and is a frequently used therapeutic
regimen for hospitalized patients encompassing the
entire spectrum of patient population and disease
(Arnold, Elliott and Holmes, 1977). Infusion
thrombophlebitis (ITP) is the most common com-
plication of i.v. infusion (Curry and Zallen, 1973;
Arnold, Elliott and Holmes, 1977; Eremin and Mar-
shall, 1977) and is characterized by a painful local
reaction often accompanied by erythema and
oedema (Chamberland, Lyons and Brock, 1977).
Symptoms and signs usually lasts days or weeks,
although Hastbacka and colleagues (1965) reported
that symptoms may persist for months. It is possible
for suppuration (Ross, 1972; Curry and Zallen,
1973; Arnold, Elliott and Holmes, 1977), septi-
caemia (Arnold, Elliott and Holmes, 1977) and
rarely pulmonary embolism (Swanson and Aldrete,
1969) or death (Frazer, Eke and Laing, 1977) to
result.
Warthen (1930) was perhaps the first to describe
the condition: "About the third or fourth day the
vein wall becomes oedematous and painful from the
constant flow of dextrose, which is mildly irritating.
The lumen of the vessel is decreased by this oedema
and the flow gradually diminishes." This succinct
description remains as true today as it was over 50
years ago.
CLINICAL DEFINITION
There is a spectrum of definitions in the literature as Erythema around the cannula site, but not extend-
illustrated by the following:
(1) "Redness and tenderness and oedema of the
vein" (Medical Research Council subcommittee,
G. B. H. LEWIS, M.B., B.S., M.LITT., DA., F.F.A.R.A.C.S.; J. F. to less than five cms. proximal to the tip of the can-
HECKER, B.SC, B.V.SC, PH.D.; Department of Physiology, The nula."
University of New England, Armidale, 2350 New South Wales,
Australia.
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
SUMMARY
Infusion thrombophlebitis is a common complica-
tion of i.v. infusions. Many factors appear to be
involved in its aetiology, of which the duration of
infusion, the drugs infused and the solution(s)
infused are the most important. Effective
prophylaxis should be based on an understanding
of the possible pathophysiology.
1957; Skajaaetal., 1961;Thayssen, 1973;Hessovet
al., 1977).
(2) "Tenderness and/or erythema along the vein,
incurred up to 14 days after the infusion"
(Hastbacka etal., 1965).
(3) "The polyethylene catheter was removed when a
site was no longer judged to be viable because of
pain, tenderness, oedema or inability to receive the
prescribed fluids." (Daniell, 1973—who did not
diagnose thrombophlebitis before this stage).
(4) Dinley (1976) modified the (British) Medical
Research Council (1957) criteria as follows:
Grade 0 "No reaction or discomfort, puncture
wound clean."
Grade 1 "Induration around the vein. Slight ten-
derness may be present at the infusion site. No pain
on speeding up the infusion rate and no evidence of
phlebitis."
Grade 2 "Mild discomfort at the infusion site, ten-
derness, over cannula and just proximal to it. Slight
discomfort on increasing the infusion rate.
ing beyond the tip of the cannula."
Grade 3 "Moderate discomfort at the intravenous
site. Constant moderate to severe pain when the
intravenous rate is increased. Erythema extending
Grade 4 "Moderate to severe discomfort at the
INFUSION THROMBOPHLEBITIS 221

TABLE I. Incidence of infusion thrombophlebitis

Year Author(s) Incidence (%) Comment

1951 Bolton-Carter 52
1952 Handfield-Jones and Lewis 68
1952 Page, Raine and Jones 65
1954 Jones 73
1957 Medical Research Council 56
1959 McNair and Dudley 100 Long saphenous veins
1961 Skajaa and colleagues 71
1965 Hastbacka and colleagues 25 Of 1048 infusions
1966 Elfving and Saikku 18
1967 Kay and Roberts 17
1969 Swanson and Aldrete 7.4-47.9 Depending on needle size
125
1972 Pussell and Pitney 75 I detection

Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017

1972 Ross 35 Of 35 infusions
1975 Collin and colleagues 43.2 Over 72 h
1976 Dinley 49.4
1976 Stephen and colleagues 31
1977 Arnold, Elliott and Holmes 12
1977 Chamberland, Lyons and Brock 63
1977 Eremin and Marshall 43 Without buffering pH
1977 Frazer, Eke and Laing 4.3 at 24 h;
100 at 5 days
1977 Hessov and colleagues 6.1
1981 Boon and colleagues 36

intravenous site. Infusion usually markedly slowed 2 weeks and all diagnosed cases followed until
or ceased spontaneously. Erythema greater than five symptoms have disappeared, a false impression of
cms. proximal to the tip of the cannula." the incidence, and especially severity and duration,
Grade 5 "As for grade 4, but pus seen at the infu- may be gained. Many patients subsequently fail to
sion site on removal of the cannula." report any symptoms which develop. Patients may
The accumulation of 125I-labelled fibrinogen, sometimes only admit to symptoms suggestive of
which is an accurate and sensitive method of detect- mild ITP on specific questioning. These patients
ing leg vein thrombosis (Kakkar et al., 1969) was may state that they regard the development of a ten-
used by Pussell and Pitney (1972) to detect i.v. can- der swelling in the region of their infusion site as "to
nula thrombosis 24—48 h before any clinical evi- be expected". There may be many mild cases of ITP
dence of reactions such as tenderness, inflammation
and swelling. This more objective method of diag- TABLE II. Duration of infusion thrombophlebitis
nosing thrombosis showed an overall incidence of
75% at i.v. cannula sites, but suffers from the disad-
Series Duration
vantage that the development of ITP may not neces-
sarily involve thrombosis. Gjares(1957) 49% < 1 week
The most comprehensive of these definitions is 44% > 2 weeks
30% > 1 month
that of Dinley (1976). It suffers from the disadvan- 1%> 6 months
tage that not all of the signs may develop or develop
in the sequence indicated. For ward use, the first Skajaa and colleagues May persist for many weeks
(1961)
two are possibly preferable.
Eerola and Pontinen Average 53 days
(1964)
INCIDENCE
Hastbacka and colleagues Average 28 days
The incidence of ITP varies widely (table I), reflect- (1965) Maximum 7 months
ing the lack of a standard clinical definition. In addi- 15% < 2 weeks
tion, the reported duration of phlebitic signs varies Pederson May remain for some days or
widely (table II). (1970) weeks, thus prolonging
Unless the infusion site is examined regularly for convalescence
222
which last for a short time and are never reported.
Experience with more than 10 000 infusions
suggests that these venous reactions would corres-
pond to Grade 1 or Grade 2 used in Dinley's (1976)
series.
PATHOLOGY
Pathophysiology
Rudolf Virchow (cited by Little, Loewenthal and
Mansfield, 1974) suggested, more than 100 years
ago, that thrombus formation might result from:
(1) changes in the blood,
(2) changes in the characteristics of the flow of the
blood,
(3) changes in the vessel wall.
These postulates continue to form the basis of
modern thinking, with particular emphasis on the
possible changes in the blood (Little, Loewenthal
and Mansfield, 1974).
It has been suggested that chemical irritation
causes much of the inflammation in ITP (Jones,
1954; Eerola and Pontinen, 1964; Hastbacka et al.,
1965; Elfving and Saikku, 1966; Fonkalsrud, Mur-
phy and Smith, 1968). Prostaglandins may be
released or stimulated by histamine, adenosine-5-
triphosphate (ATP) and prostaglandin E] during the
inflammatory response (Chahl and Chahl, 1976).
Prostaglandins PGE[ and PGE2 increase local vas-
cular permeability via histamine and 5-hydroxtrypt-
amine in rat and man (Crunkhorn and Willis,
1971). The early histological changes observed by
Ghildyal, Pande and Misra (1975), namely swelling
of endothelial cells and polymorphonuclear (PMN)
leucocytic infiltration to the tunica media, could
perhaps be initiated by the metabolites of
arachidonic acid. When the vein wall is injured,
platelet phospholipase A2 is stimulated to form
biologically active compounds PGE2 and PGD2
which increase vascular permeability and hydroxy-
eicosatetraenoic acid (HETE) and thromb-
oxane B2 (TXB2) which are chemotactic for PMN
leucocytes (Moncada and Vane, 1978). PGG2, PGH2
and thromboxane A2 are other platelet-produced
arachidonic acid metabolites which are active as
platelet aggregating or releasing agents and which
may also be involved at the endothelial surface.
Humoral agents released in response to venous
irritation may provoke venoconstriction. If the flow
of blood in the vein is diminished, irritant infusates
are not rapidly diluted with blood and this, together
with stasis, would predispose to ITP. Sharpey-
BRITISH JOURNAL OF ANAESTHESIA
Schafer and Ginsburg (1962) studied the effects of
humoral agents on venous tone. Adrenaline, nor-
adrenaline, 5-hydroxytryptamine and histamine all
increased venous tone in the forearm. Nitrite
decreased tone whilst increasing forearm blood
flow. Mason and Braunwald (1965) observed the
effect of sublingual nitroglycerine on the human
forearm venous tone. Forearm blood flow was
increased whilst forearm vascular resistance and
venous tone were decreased.
It is possible that the local application of nitro-
glycerine ointment near the site of an i.v. infusion
may prevent or reduce the effects of released
humoral venoconstricting substances, increase
blood flow and prevent or delay the onset of ITP.
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
Hecker, Lewis and Stanley (1983) have
demonstrated the effectiveness of very small appli-
cations of nitroglycerine ointment on the dorsum of
the hand as a local venodilator before venepuncture.
Nitroglycerine stimulates the synthesis of prostacyc-
lin by cultured human endothelial cells (Levin et al.,
1981). Prostacyclin is a very potent inhibitor of
platelet aggregation and is synthesized in vascular
walls by prostaglandin endoperoxides (Gryglewski
etal., 1976). It was suggested by Moncada and Vane
(1977) that, when a vessel is injured, endothelial
damage could reduce prostacyclin synthesis, result-
ing in regional vasoconstriction. The potent vaso-
constrictor thromboxane A2 is then produced by
aggregating platelets in the damaged area. Prosta-
cyclin, which also relaxes smooth muscle cells in
vein walls, may be produced in increased amounts by
surrounding undamaged endothelial cells. The
maintenance of effective prostacyclin concentra-
tions in veins where infusions are sited may help to
relax smooth muscle, preventing venoconstriction
and platelet aggregation and possibly ITP.
The leukotrienes are a newly recognized group of
biologically active mediators derived from
arachidonic acid. There are several reasons why it is
logical to suggest that these interesting compounds
may be involved in the development of ITP.
Leukotrienes may be produced by incubating
arachidonic acid with polymorphonuclear leuco-
cytes (Samuelsson, 1980) and these cells infiltrate
vein walls as ITP develops. Leukotrienes possess
potent pharmacological actions on smooth muscle
and could possibly affect venous tone and the flow
rates of infusion. In a recent study (Denis et al.,
1982) leukotrienes were found to increase vascular
permeability in the skin of guineapigs—relevant, as
oedema is a prominent sign in clinical ITP.
INFUSION THROMBOPHLEBITIS
Thrombi can form in any area of the cardiovascu-
lar system, the site influencing the thrombus size,
composition and shape. Sodeman and Sodeman
(1967) state that in phlebothrombosis the thrombus
results from slowing of the blood stream and
increased coagulability of the blood rather than from
inflammation, whereas thrombophlebitis results
from inflammation of the venous wall. They note
that not all agree that phlebothrombosis and throm-
bophlebitis are sharply defined entities. Little, Loe-
wenthal and Mansfield (1974), describing superfi-
cial phlebitis, make the following observations. The
superficial veins of the limb are affected, there is
thrombus within the lumen of the vein and an acute
inflammatory reaction in the vein wall and in the
surrounding tissues which results in pain.
Perhaps the frequently sketchy coverage accorded
ITP in surgical and anaesthetic texts reflects the
relative paucity of sound scientific papers on this
subject. This review may assist to demonstrate the
high incidence, inadequately understood patho-
genesis, continuing morbidity and relative failure
of prophylactic and therapeutic measures cur-
rently used in this iatrogenic condition.
Even less clear is the relationship of the pathology
if ITP to the other frequent problem with infusions,
that of extravasation of infusate ("tissuing").
Histopathology of superficial thrombophlebitis
Histopathological study of human veins following
ITP has been limited (Thomas', Evers and Racz,
1970). A small piece of vein was removed under local
anaesthesia from each of 50 patients with ITP by
Ghildyal, Pande and Misra (1975), who also studied
bacteriologically the tips of catheters, needles and
other infusion devices. Histological changes were
categorized as mild, moderate or severe. Mild
changes in 23 specimens consisted of swelling of
endothelial cells and PMN leucocyte infiltration in
the tunica media. A further 18 specimens displayed
cellular destruction and cellular breakdown of the
endothelium. Oedema, PMN leucocyte infiltration
and pyknotic nuclei of muscle cells were seen in the
tunica media. The remaining nine were severely
affected, with destruction of endothelium, cell nuc-
lei pyknosis in the media with oedema and PMN
leucocytic infiltration. Additionally, there was
haemorrhage and necrosis of the wall. The damage
was most severe when a thrombus was present. Five
positive cultures of catheters or i.v. devices grew
either Staphylococcus aureus (two), Streptococcus
haemolyticus (two) or Pseudomonas pyocyanea (one).
223
The authors concluded that there was little or no
correlation between the bacterial cultures and
phlebitis in their study. The changes in the vein
were similar to earlier observations, including those
of Gritsch and Ballinger (1959), who studied tissue
reactions of dogs in which i.v. plastic tubing had
been implanted.
Animal studies of ITP are also limited. Horvitz,
Sachar and Elman (1943), who studied the veins of
11 dogs after infusions of 5% and 10% glucose solu-
tions, found vacuolation of endothelial cells, fol-
lowed by progressive cellular breakdown, leading to
death of cells and destruction of the lining. Fibrin
and thrombus were often deposited on the denuded
subintima.
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
ITP was induced in dogs by i.v. infusion of
dextrose, saline to which, in some cases, repeated
doses of naftidrofuryl (Praxilene) had been added by
Woodhouse (1980). The infusions were continued
for up to 12 h before sacrifice of the dogs. Histolog-
ical examination showed early changes after 6 h in
all specimens. More severe changes were noted in
dogs sacrificed later. The infusions to which naftid-
rofuryl had been added produced only marginally
more severe and earlier changes than dextrose/saline
control infusions. In the early stages, histological
examination showed fibrin and PMN leucocyte
adherence to the endothelium of the vein. As the
severity increased, leucocyte infiltration through the
endothelium occurred, progressively affecting the
deeper layers to the adventitae. Thrombosis was
noted in three of 16 veins. Woodhouse concluded
that the essential event in ITP is infiltration of the
vein wall by circulating leucocytes.
Woodhouse (1980) and Ghildyal, Pande and
Misra (1975) are in agreement that ITP is usually an
acute sterile inflammation and the early study of
Horvitz is consistent with this view.
Suppurative thrombophlebitis
There is a separate condition of suppurative
thrombophlebitis which may be associated with cut-
down infusions (Woodhouse, 1980). A brief review
of the features of this serious complication is indi-
cated, to define more clearly the differences between
the two conditions. It is possible a previously sterile
ITP may become infected and develop into suppura-
tive thrombophlebitis.
Stein and Pruitt (1970) studied 295 patients,
admitted to hospital with burns, in whom venous
cannulations were performed by cut-down. Sup-
purative thrombophlebitis developed in 24 patients,
224 BRITISH JOURNAL OF ANAESTHESIA

in only 11 of whom was the diagnosis made before
death. Fever was present in 10 of the 11. Six patients
had exudation of purulent material from the vein.
Pain, tenderness, a red streak and swelling, common
findings in ITP, were rare in this series. In 14
patients no excision was performed; 13 were diag-
nosed at autopsy and the one case treated conserva-
tively with antibiotics, elevation, warm compresses
and dextran, died of septicaemia. Of the other 10
patients who were treated by extensive excision of
the involved vein, seven survived. If the diagnosis is
in doubt, exploratory venotomy is recommended.
Stein and Pruitt (1970) claimed that this condition
was responsible for 2.4% of deaths. They proposed
the following sequence to explain the clinical and
histological findings. Firstly, an i.v. cannula is sited
in a vein and a fibrin clot forms on the vein wall or
catheter tip. Subsequently, micro-organisms are
trapped in the clot when they enter the vein via the
cannula or venous tributaries. Finally, the infected
clot serves to seed the blood-stream with.micro-
organisms. Patients with thermal injury are notori-
ously prone to infection, and scrupulous attention to
asepsis is constantly emphasized, but any evidence
suggestive of suppuration associated with i.v.
therapy in a patient is necessarily viewed with con-

Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017

TABLE III. Duration of infusion

Duration of infusion (h) and ITP

Series and
year 12 24 48 96 192
Bolton-Carter < I
(1951) 4.5% I- -> 52%
Bogen(1960)
0.9% 37.5%
Cheney and -I
Lincoln (1964) 9% I -> 58%
Eerola and -> 25.9% overall incidence av. 4—5 h
P6ntinen(1964)
Hastbacka and < I
others(1965) 62% of ITP occurred in first week
Kay and Roberts < I
(1967) (low) | - > (high)
Swanson and <- - - - - -I
Aldrete(1969) 20% of ITP I 68%ofITP
Brown (1970) < - I
8% I > 18%
Ross (1972) "Duration significantly related to ITP (P < 0.001)"
Collin and 61% (plastic cannulae) I >
others(1975) >
47% (scalp vein needles) I
Dinley(1976)
"Duration related to incidence and severity"
Ferguson and
others (1976) (0% with heparin lock needles in situ)
(30% with heparin lock needles in situ) I
Stephen and "Duration of cannulation correlates with development
others (1976) of phlebitis in a linear fashion"
Chamberland, - I
Lyons and "most patients had inflammation"
Brock(1977)
Eremin and -I 10%
Marshall (1977) -I 48%
-I 80%
Frazer, Eke and -I 4.3%
Laing(1977) -I 100%
INFUSION THROMBOPHLEBITIS 225

cern. the antecubital fossa is utilized.

PREDISPOSING FACTORS
The aetiology of ITP appears to be a combination of
factors (Curry and Zallen, 1973).
Duration of infusion
The duration of infusion is such an important fac-
tor that undoubtedly it influences the effect of other
factors (Ross, 1972). Nearly all investigators have
reported that duration is significant (table III).
Site of infusion and size of vein
Infection
Although any thrombus which develops around a
cannula in a vein would seemingly provide an ideal
site for bacterial multiplication, infection would not
appear to be a major cause of ITP in the upper limb.
Many of the studies cited in table V fail to show any
correlation between infection and ITP. If the cause
were bacterial, one might expect that antibiotic
therapy would reduce the incidence of ITP, but
there does not appear to be any study that
demonstrates this. This does not exclude the possi-
bility that ITP and infection may occur together,
the infection being amenable to antibiotic therapy.

Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017

Data on the incidence of ITP related to the site of
infusion are summarized in table IV. There is a ten-
dency for smaller veins such as those on the dorsum
of the hand to have a higher incidence of ITP. How-
ever, the series do not appear to be consistent and
one may conclude that there is no clear evidence that
any particular commonly used upper limb vein is
more likely to develop ITP.
Few authors have commented on vein size. Those
that do comment, agree that larger veins are to be
preferred (Jones, 1957; Swanson and Aldrete, 1969;
Sketch etal., 1972; Eremin and Marshall, 1977), but
many patients prefer to have an i.v. infusion in the
wrist or hand; for this is less restrictive than when
TABLEIV. Site of infusion
Types of solution infused
Most i.v. solutions have a low pH. Thus, solu-
tions containing dextrose are usually of pH 3.4—5,
while other solutions are around pH 5-6 (Tse and
Lee, 1971). Reasons given for manufacturing such
acid solutions are to prevent caramellization of
dextrose (Ross, 1972), and to prevent reaction with
glass containers during autoclaving. Despite the low
pH, the buffering capacity of solutions is low.
Amino acid solutions are less irritant to veins
when the pH is adjusted to 7.4 (Horvitz, Sachar and
Elman, 1943): A number of workers have investi-

Incidence at infusion site (%)

Dorsum Dorsum Ante-
Series and of of Forearm cubital Cephalic Other
Year hand wrist vein fossa vein

Gjares(1957) 40 — 9
Eerolaand Pontinen
(1964) 41.8 — _ 26.1
Fonkalsrud, Murphy
and Smith (1968) 18.1 — 23.4 _
Swanson and Aldrete
(1969) 34.6 32.5 19.4 33.3
Brown (1970) 24 — 30 —
Eremin and Marshall Subclavian
(1977) 73 — 65 51 63 12
Skajaa and others
(1961) No significant differences
Thomas, Evers and Venous complications lowest in the dorsum of the hand but not
Racz(1970) significant when time and cannulae considered
Ross (1972) More proximal veins (forearm) — higher incidence than distally
Boon and others (1981) No significant differences
226
gated i.v. dextrose solutions. Vere, Sykes and
Armitage (1960) conducted a blind, controlled trial
of dextrose solutions sterilized by autoclaving or by
filtration, and found that autoclaved solutions were
acid and were associated with a significantly higher
incidence of ITP than were filtered solutions of pH
near neutrality. Buffering of solutions just
before use has been demonstrated to reduce the inci-
dence of ITP in both short- (Fonkalsrud etal., 1971)
and longer-term infusions (Eremin and Marshall,
1977).
TABLE V. Comments on infection
Series Comment
Jones(1957) Argued against infection causing ITP,
stating that "infection is never seen".
Cheney and Only 9.7% of cases with ITP grew
Lincoln (1964) pathogenic bacteria from equipment.
Kay and Roberts One of 16 blood cultures positive.
(1967) Two of 23 catheters had positive cultures.
Banks and others 45% of catheter tips grew bacteria, but only
(1970) four of 118 patients had bacteraemia from
same catheter/skin organism. No relation
to duration, fluid or catheter type.
Collin and others Teflon cannulae significantly associated
(1975) with bacterial contamination and ITP.
Ferguson and Positive flush culture correlated with
others (1976) phlebitis when heparin lock needles used.
Stephen and others No instances of septicaemia despite a
(1976) positive bacterial culture rate from
cannulae of 48% of 130 patients.
Eremin and 22% of i.v. cannulae grew bacteria whether
Marshall (1977) or not there was ITP or other infusion
complication.
Archer and Fowler No evidence that special skin preparation
(1977) decreased incidence of ITP.
The low pH of i.v. fluid preparations has been
implicated by other authors (Elfving and Saikku,
1966; Fonkalsrud, Murphy and Smith, 1968; Tse
and Lee, 1971). Gritsch and Ballinger (1959)
demonstrated in dogs that great damage to all layers
of the vein wall was produced by acid solutions.
Stephen and colleagues (1976) found that there was
a significantly smaller incidence of ITP when nor-
mal saline alone was given, compared with other
isotonic fluids or normal saline in combination with
other fluids. Vascular endothelium is particularly
sensitive to pH and in animal experiments platelet
adherence to the damaged surface occurs with leuco-
cyte infiltration and oedema (Eremin and Marshall,
1977). Maximal damage is seen near the tip of the
cannula, but can spread proximally and cause
BRITISH JOURNAL OF ANAESTHESIA
intraluminal thrombosis. Prolonged infusion aug-
ments these pathological changes, but they are
diminished by perfusion of solutions of neutral pH
(Eremin and Marshall, 1977). Although alteration of
pH towards neutrality may decrease venous irrita-
tion and ITP, Williams and Moravec (1967)
reported that changing the pH of a solution can
change the efficacy of some added drugs.
Hypertonic solutions also irritate veins (Gritsch
and Ballinger, 1959). Since amino acid solutions
and 50% dextrose can be tolerated only if given at a
slow rate in large veins which have a high blood flow
(Ravitch, 1969; Wilmore and Dudrick, 1969), they
are usually infused through a central line.
Skajaa and co-workers (1961) found no difference
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
in the incidence of ITP between 5% dextrose,
physiological saline and blood. However, Dinley
(1976) reported an extremely high incidence of ven-
ous reactions to blood transfusion. He commented
that, as the vast majority of these transfusions were
given from bottled blood, the plasticizers of
polyvinylchloride bags can be excluded as a cause.
Paniculate matter
Extraneous, mobile, undissolved substances in
parenteral solutions constitute particulate matter.
Examples are rubber, chemicals, glass, cellulose
fibres and fungi (Garvan and Gunner, 1963).
Ryan, Rapp and De Luca, (1973) found that the
use of a 0.45-um filter reduced the incidence of
phlebitis in patients receiving infusions for 72 h. De
Luca and associates (1975) also found ITP to be sig-
nificantly reduced in patients after operation when
an in-line 0.45-um filter was used. Turco and Davis
(1971) and Davis and Turco (1973) have shown that
every solution of every company tested contained
particulate matter 5 urn and larger. Evans, Barker
and Simone (1976) tested the effectiveness of a 5-um
filter in preventing phlebitis and found a signific-
antly lower incidence in patients receiving filtered
i.v. solutions.
Particulate irritation to the endothelium may result
in inflammation of the vein. However, some
investigators have failed to demonstrate the value of
in-line filtration in reducing the incidence of ITP.
Collin and colleagues (1973) and Chamberland, Lyons
and Brock (1977) point out that the flow of infusate
with 0.45-um filters is impeded, filters need to be
changed perhaps daily, and there is added expense.
Collin and co-workers (1973) observed an incidence of
phlebitis of 40% with filters and 47% without.
Chamberland, Lyons and Brock (1977) found a 63%
INFUSION THROMBOPHLEBITIS
incidence with filters, and 58% of patients without
filters developed ITP.
In-line filtration, perhaps because of added cost,
extra work and a restriction in the flow capacity of the
infusion with small filters, is not commonly practised,
although in theory it is a good idea, and there is some
experimental evidence that it may be worthwhile.
Equipment
Needle or cannula. Skajaa and colleagues (1961)
used three needle sizes, but did not observe differ-
ences in the incidence of ITP. Swanson and Aldrete
(1969) found a significantly higher incidence with
larger catheters than with 19-gauge needles. Curry
and Zallen (1973) recommended the use of small-
gauge (18-20) catheters and large diameter veins to
reduce ITP. Thrombus formation is related to the
size of device (Spanos and Hecker, 1976; Hecker,
1980), and so a build up of thrombus on a small can-
nula might be less likely to occlude the vessel. Even
a partial occlusion may impede blood flow and a
deleterious effect on the endothelium may be pro-
duced by the undiluted infusate (Hecker, 1980).
Winged Teflon cannulae may restrict movement
in all planes, reduce irritation of the intima and
reduce ITP.
Cannula material. Dinley (1976) showed that the
incidence of venous reactions was related to the can-
nula material. He used cannulae of four types of
material. Fluoroethylene propylene Teflon can-
nulae caused less reactions than tetrafluoroethylene
Teflon. Polyvinylchloride and polyethylene can-
nulae were more irritant than either type of Teflon.
Hecker (1980) measured thrombus formation
associated with seven types of cannulae made of
polyethylene, polypropylene and types of Teflon
and found a significant difference of greater than
two-fold in the amount of thrombus formed.
The roughness of the tubing or catheter tip may
vary with the extrusion temperature during man-
ufacture. Hecker and Edwards (1981) tested a
polyvinylchloride resin tubing extruded at five diff-
erent temperatures for thrombogenicity. Lower
extrusion temperatures were associated with greater
roughness and more thrombus formation.
Trauma at venepuncture. James (1954) suggested
that greater experience in performing venepuncture
decreases the incidence of ITP. Watt (1977) was
concerned that the trauma of insertion provides
bacterial access to the circulation. Eremin and Mar-
shall (1977) suggested that trauma is more likely to
occur to smaller veins at the time of insertion of the
227
cannula and they felt that plastic cannulae inserted
through an i.v. needle may be associated with less
trauma to the vein than the plastic cannulae inserted
as a sheath around the venepuncture needle. They
found that complications were highest with the lat-
ter type of cannula and, also, highest in small veins.
In other studies, Jones (1957) found that vene-
puncture was not a significant factor and McNair
and Dudley (1959) regarded it as a minor factor.
Hastbacka and associates (1965) reported that the
incidence of ITP did not seem to be any higher when
a haematoma formed as a result of venepuncture.
Infusion set. Handfield-Jones and Lewis (1952)
obtained a reduction of 50% in the frequency of ITP
when they used plastic instead of rubber sets. This
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
finding was confirmed by the (British) Medical
Research Council (1957). Rubber tubing is of his-
toric interest, but more recently Skajaa and col-
leagues (1961) found no difference between two
types of plastic infusion sets.
Rate of flow
Hessov and Bojsen-M0ller (1976) suggest that the
rate of infusion is important, with it being advan-
tageous to give irritant solutions quickly. Ross
(1972) could find no studies which compared infu-
sion flow rates with the incidence of phlebitis. In her
series, the mean infusion rate ranged from 59 to
325 ml per hour. She felt that a significant correla-
tion between slower end-infusion rates and the inci-
dence of phlebitis may indicate that injury to the
vein by an irritating solution is apt to occur with slow
infusions. It is possible that the lower end-infusion
rates were the result of venoconstriction induced by
the infusion and that this may be an early indication
of the likely development of ITP. Evans, Barker and
Simone (1976) found no correlation between flow
rate and the incidence of phlebitis.
Age
Aldman and Garsten (1960), Hastbacka and col-
leagues (1965), Fonkalsrud, Murphy and Smith
(1968), Swanson and Aldrete (1969), Dinley (1976)
and Archer and Fowler (1977) were unable to show
a consistent trend between age and the incidence of
ITP. Hastbacka and co-workers (1965) showed a
tendency for the incidence to be lower in younger
age groups, but Skajaa and colleagues (1961) found
it was the younger patients who tended to have the
higher incidence.
228
Sex
Hastbacka and co-workers (1965) noted that the
frequency of phlebitis in females was about twice as
high as in males after thiopentone injection: after
pethidine injection no difference was seen. Archer
and Fowler (1977) reported that the female inci-
dence was twice that of males. No sex-related differ-
ences were detected by Aldman and Garsten (1960),
Skajaa and associates (1961) or Dinley (1976).
Administered drugs
Benzodiazepines. Wyant and Studney (1970)
observed that phlebitis was not uncommon when
diazepam was administered to induce anaesthesia,
but the incidence of venous reactions after i.v. benzo-
diazepines varies in different studies. This may be
attributable to solvents, size of veins, method of
injection and follow up. Hegarty and Dundee (1977)
reported that age may also be important. A time-
related increase in frequency during the first days
after injection was noted by Siebke, Ellertsen and
Lind (1976), Hegarty and Dundee (1977) and Mik-
kelsen and colleagues (1980). Langdon, Harlan and
Bailey (1973) observed an ITP incidence of only
3.5% after i.v. diazepam, whereas Hegarty and
Dundee (1977) found an incidence of 39%. Mik-
kelsen and co-workers (1980) reported an incidence
of about 14% after both diazepam and flunit-
razepam, whereas Hegarty and Dundee (1977)
observed only a 5% incidence with nitrazepam. The
solvent used in the preparation of the parenteral
benzodiazepines such as flunitrazepam and
diazepam is commonly propylene glycol, which may
play a significant role in the aetiology. When
oxazepam dissolved in propylene glycol was injected
i.v. to dogs (Alvan et al.,1974), haemolysis and ITP
were observed after rapid infusion and were shown
to be caused by the properties of the vehicle. Mat-
tila, Rossi and Ruoppi (1981) showed a reduction of
venous sequelae of i.v. diazepam when a fat emul-
sion was used as the solvent.
Barbiturates and other anaesthetic agents. Barbitu-
rates have been implicated as aetiological agents by
Kivalo and Tammisto (1959), Eerola and Pontinen
(1964), Hastbacka and colleagues (1965), Hewitt
and co-workers (1966), O'Donnell, Hewitt and
Dundee (1969), Carson and colleagues (1972) and
Jamieson. Desjardins and Caron (1972). Hastbacka
and associates (1965) found that complications
occurred more often after the injection of 5% sodium
thiopentone than after 2.5% solution.
BRITISH JOURNAL OF ANAESTHESIA
Hewitt and colleagues (1966) reported that the
injection of propanidid was associated with an
increased incidence of ITP. O'Donnell, Hewitt and
Dundee (1969) noted that greater dilution of pro-
panidid significantly decreased the incidence of ven-
ous sequelae. Carson and colleagues (1972) reported
that one of the major drawbacks in the development
of non-barbiturate i.v. anaesthetic agents has been
the high incidence of local venous thrombosis. They
cited hydroxydione as an example. They found that
the newer steroid Althesin had a lower incidence
than either 5% propanidid or 5% thiopentone solu-
tions. There was no difference between the effects
of Althesin, 2.5% thiopentone and 1.0%
methohexitone.
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
Etomidate causes such severe pain to some
patients on injection that its value for induction is
limited (Kenny, 1981). Pain on injection was
avoided by Lees and colleagues (1981) who used
etomidate and fentanyl only as an infusion for
maintenance of anaesthesia, but they reported a 22%
incidence of thrombosis or ITP compared with a
10% incidence in control patients in whom
halothane was the maintenance agent.
Boon and co-workers (1981), in a study of 16
drugs commonly used in anaesthetic practice, con-
cluded that "the injection of anaesthetic agents into
an intravenous line does not result in an increased
incidence of local thrombosis or phlebitis in the 48
hour period after surgery". However, the incidence
of ITP in their series was 36%.
Antibiotics. "Antibiotics may also be associated
with phlebitis and other routes of administration
should be considered" (Stephen et al., 1976).
Cephalosporin antibiotics have been associated
with a high incidence of ITP (Perkins and Saslaw,
1966; Storey, 1971; Bran, Levison and Kay, 1972;
Lane, Taggart and lies, 1972; Inagaki and Bodey,
1973). Siebert and colleagues (1976) studied the
incidence and severity of ITP associated with
cefamandole, cephapirin and cephalothin. The inci-
dence was similar for each drug, but ITP was sig-
nificantly more severe with cephalothin than the
other agents. These results are in agreement with the
findings of Lane, Taggart and lies (1972) and
Inagaki and Bodey (1973).
Tetracyclines added to i.v. infusions had no effect
on the complication rate in Eremin and Marshall's
(1977) study, but increased the risk of phlebitis in
Kay and Roberts' (1967) series. Crystalline penicil-
lin and ampicillin had no effect on the incidence
when added to Eremin and Marshall's (1977) infu-
sions. Addition of ampicillin to infusions reduced
INFUSION THROMBOPHLEBITIS
the incidence of ITP in Brown's (1970) study, but
Thomas, Evers and Racz (1970) gained a strong clin-
ical impression that the onset of phlebitis was has-
tened by ampicillin.
Other drugs. Kay and Roberts (1967), Thomas,
Evers and Racz (1970) and Stephen and colleagues
(1976) all found that potassium chloride added to an
infusion increased the risk of phlebitis, but Eremin
and Marshall (1977) found no change in risk. Nor-
dell, Mogensen and Nyquist (1972) found a correla-
tion between i.v. lignocaine infusions and ITP.
Hastbacka and co-workers (1965) reported that
injection of undiluted pethidine increased the inci-
dence of ITP by about one-third. A histamine-like
reaction along the course of the vein followed
pethidine administration in 25% of cases. The inci-
dence of this reaction showed a correlation with sub-
sequent ITP, contrary to the situation with barbitu-
rates, with which ITP does not appear to be related
to an acute venous response.
Other factors
A history of previous phlebitis or allergy have been
suggested as possible causative factors, but no corre-
lation between these factors and ITP has been found
(Aldman and Garsten, 1960; Hastback et al., 1965).
Brown (1970) commented that, if a needle is left in a
vein without an infusion, the incidence of ITP is
very low, even after 72 h. Obesity does not seem to
be a factor (Hastbacka et al., 1965).
Michaels and Reubner (1953) suggested that con-
taminated infusion fluid may give rise to phlebitis,
but this should be considered under the heading of
suppurative thrombophlebitis discussed earlier.
Although cardiovascular disease did not appear to
predispose to ITP, Hastbacka and colleagues (1965)
found an incidence of 53% in those patients with a
circulation time of greater than 19 s, compared with
35% with a circulation time of less than 15 s. A
further analysis of those whose circulation time
exceeded 25 s showed a phlebitis incidence of 88%.
No other studies of a similar nature appear to have
been reported.
TREATMENT OF INFUSION THROMBOPHLEBITIS
Removal of the i.v. device at the earliest opportunity
is the primary treatment. Since ITP is a sterile
inflammation, the use of anti-inflammatory agents is
indicated. Oxvphenbutazone has been used to treat
ITP. Archer and Fowler (1977) point out that
phenylbutazone and oxyphenbutazone became
accepted as therapeutic agents without a solitary
229
satisfactory double-blind controlled study. They
agree that their study is also flawed, for when study-
ing the efficacy of oxyphenbutazone, the treatment
group were initially more affected by ITP than the
controls with regard to pain, area of erythema and .
length of indurated vein. Despite bias against the
treatment, they found oxyphenbutazone effective in
rapidly relieving pain and tenderness. A 70%
improvement in pain when compared with initial
severity occurred within 4 days, compared with a
43% improvement in the placebo group. A parallel
improvement in tenderness was also noted. The area
of erythema showed marked resolution in the same
period. They concluded that the reduction in pain
occurred with resolution of inflammation and was
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
not merely the result of an analgesic effect.
Application of other anti-inflammatory agents
such as hydrocortisone cream is common and
analgesics are given for symptomatic relief.
Hirudoid, a heparinoid preparation has been
applied, but Hastbacka and colleagues (1965) stated
that the value of this treatment seems questionable.
In their series, the symptoms of ITP sometimes per-
sisted for months, the longest for 7 months after
infusion.
ITP is a painful condition and therefore analgesics
are indicated. Drugs such as aspirin may inhibit
pathological changes associated with prostaglandins
(see next section).
PREVENTION
The surest way to prevent ITP is to avoid setting up
or continuing an infusion unnecessarily. When pro-
longed i.v. therapy is required, it is recommended
that the site of infusion be changed every 24 or 48 h
where practical, as duration of infusion is an impor-
tant factor. This is not always practical.
Another way to reduce the incidence of ITP is by
adding certain drugs to the infusion solution.
Anderson (1951) showed a significant delay in
the development of ITP with the addition of
heparin. Daniell (1973) added heparin lOOOi.u./litre
5% Dextrose B.P. and reduced the incidence of ITP
in a double-blind controlled study. Other studies
have also demonstrated a reduction in thrombosis
when heparin is given (Martin, 1944; Erwin, Strick-
ler and Rice, 1953; Wessler and Rogers, 1956; Hohn
1966; Sketch et al., 1972; Stephen et al., 1976) and
Eremin and Marshall (1977) also commented that
heparin appeared to reduce the incidence, although
their numbers were small. A number of studies
utilizing different types of catheter (McNair and
Dudley, 1959; Martin, 1965; Roy, Wilkinson and
 230
Bayliss, 1967; Walters, Stanger and Rotem, 1972)
have not demonstrated a significant effect. Pussell
and Pitney (1972) observed that the incidence of
thrombosis at the i.v. cannula site was approxi-
mately 75% in patients who were receiving
anticoagulant drugs (heparin or heparin followed by
warfarin) and in those who were not, in a study of
patients following myocardial infarction. Polok
(1956) advocated the use of hydrocortisone and
McNair and Dudley (1959) reported that it defi-
nitely reduced the incidence and severity of ITP.
Clark, Polak and Hajnal (1960) showed that the inci-
dence decreased when hydrocortisone was added to
the infusion solution. Schafermeyer (1974) recom-
mended that heparin 5 mg and hydrocortisone 1 mg
be added to each litre of i.v. fluid when cephalothin,
gentamicin, kanamycin and potassium salts are to
be administered.
A different regimen was adopted by Woodhouse
(1979), who used a local heparin and hydrocortisone
cream over the infusion site. He found that 15 of 49
control patients, compared with only eight of 48
treated patients, developed ITP. Patients in the
treatment group took longer to develop ITP (66.4 h
cf 47.4 h). Eerola and Pontinen (1964) investigated
the prophylactic effect of two anticoagulant oint-
ments, Thrombosol forte (containing heparin) and
Hirudoid (containing a heparinoid). Following infu-
sions for an average duration of 4-5 h, 21.4% of the
arms with Hirudoid or Thrombosol forte treatment
showed signs of ITP. The corresponding figure in
the untreated group was 25.9%. The duration of
ITP, when it developed, was shorter in the treat-
ment group.
Prostaglandins and related compounds are
involved in the processes of haemostasis and throm-
bosis. Aspirin, indomethacin and phenylbutazone
inhibit prostaglandin synthesis. Thromboxane A2,
which is synthesized by platelets and other cells,
induces platelet aggregation, whereas PGI2, which is
synthesized by vascular wall cells, inhibits platelet
aggregation (Moncada et al., 1976). Aspirin inhibits
the release reaction by acetylating platelet cyclo-
oxygenase and so inhibits the synthesis of prosta-
glandins and thromboxane A2 (Roth and Majerus,
1975). Aspirin inhibits PGI2 synthesis by the vessel
wall, which could be potentially thrombogenic. The
dose of aspirin required for different effects is diffe-
rent (Kelton et al., 1978), so when aspirin is used as
an antithrombogenic agent, the doses used may
inhibit platelet prostaglandin synthesis without
affecting PGI2 formation by the vessel wall. It is now
thought that the antipyretic, analgesic and anti-
BRITISH JOURNAL OF ANAESTHESIA
inflammatory actions of aspirin-like drugs are
mediated via inhibition of prostaglandin biosyn-
thesis, but a place for aspirin in the prophylaxis of
ITP is not established.
There is little evidence that antibiotics, either by
infusion or applied locally, influence the incidence
of ITP (Norden, 1969; Zinner et al., 1969; Evans,
Barker and Simone 1976), except that ampicillin
appeared to lower the incidence in Brown's (1970)
series.
The use of buffered solutions lowers the risk of
phlebitis. Fonkalsrud, Murphy and Smith (1968)
showed that a significant reduction in endothelial
injury occurred when the infusion solution was buf-
fered to pH 7.4. Approximately 15 mmol of sodium
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
bicarbonate is sufficient to buffer 1 litre of 5%
dextrose B.P. solution.
It is common observation that infusion rates often
slow down after some hours. One likely explanation
is increased resistance brought about by local veno-
constriction at the site of infusion. This is caused
by irritation induced by infused solutions (Lewis
and Hecker, 1984) and this will decrease blood flow
past the cannula tip, reducing dilution of infusate by
the blood. Further slowing of the drip rate may
occur with continued irritation, and relative stasis
could encourage thrombus formation. Increased
exposure of endothelium to infusate might lead to
swelling of endothelial cells and development of
oedema.
CONCLUSION
The lack of a universally applied clinical definition
of ITP makes a valid comparison of incidence and
severity well nigh impossible. A multitude of factors
appears to be involved in the development of this
condition. Of those factors, the more important are
duration, drugs administered and solutions infused.
There is little detailed knowledge of the histopathol-
ogy, although there is increasing awareness of the
involvement of prostaglandins and possibly leuko-
trienes in its pathophysiology.
Prophylaxis is often inadequate, despite recom-
mendations for the neutralization of acid solutions
and the limiting of the duration of infusions. This is
reflected in the continuing high incidence, with a
significant morbidity complicating i.v. therapy.
Until this iatrogenic condition is studied under more
controlled conditions, we are not likely to under-
stand the pathophysiological details upon which
effective prophylaxis and management should be
based.
INFUSION THROMBOPHLEBITIS
REFERENCE
Aldman, B., and Garsten, P. (1960). Profylaks med Hirudoid
mot Infusionstrombophlebit. Svenska Lak Tidn., 57, 951.
Alvan, G., Jonsson, M., Sundwall, A., and Vessman, J. (1974).
First pass conjugation and enterohepatic recycling of
oxazepam in dogs; intravenous tolerance of oxazepam in pro-
pylene glycol. Ada Anaesthesiol. Scand., (Suppl.), 16.
Anderson, L. H. (1951). Venous catheterization for continuous
parenteral fluid therapy: Use of heparin in delaying throm-
bophlebitis. J . Lab. Clin. Med., 38, 585.
Archer, D. S., and Fowler, P. D. (1977). Comparison of
oxyphenbutazone and placebo in the treatment of superficial
thrombophlebitis: an object lesson in clinical trial design. Prac-
titioner, 218, 3.
Arnold, R. E., Elliott, E. K., and Holmes, B. H. (1977). The
importance of frequent examination of infusion sites in pre-
venting postinfusion phlebitis. Surg. Gynecol. Obstet., 145,19.
Banks, D. C , Yates, D. B., Cawdrey, H. M., Harries, M. G.,
and Kidner, P. H. (1970). Infection from intravenous cathet-
ers. Lancet, 1, 443.
Bogen, J. E. (1960). Local complications in 167 patients with
indwelling venous catheters. Surg. Gynecol. Obstet., 110, 112.
Bolton-Carter, J. F. (1951). Reduction in thrombophlebitis by
limiting duration of intravenous infusions. Lancet, 2, 20.
Boon, J., Beemer, G. H., Bainbridge, D. J. M., and Crankshaw,
D. P. (1981). Postinfusion thrombophlebitis: effect of
intravenous drugs used in anaesthesia. Anaesth. Intens. Care, 9,
23.
Bran, J. L., Levison, M. E., and Kaye, D. (1972). Clinical and in
vitro evaluation of cephapirin: a new cephalosporin antibiotic.
Antimicrob. Agents Chemother., 1, 35.
Brown, G. A. (1970). Infusion thrombophlebitis. Br. J. Clin.
Prac, 24, 197.
Carson, I. W., Alexander, J. P., Hewitt, J. C , and Dundee, J.
W. (1972). Clinical studies of induction agents, XLI: Venous
sequelae following the use of the steroid anaesthetic agent,
Althesin. Br.J. Anaesth., 44, 1311.
Chahl, J. S., and Chahl, L. A. (1976). The role of prostaglandins
in chemically induced inflammation. Br. J. Exp. Path., 57,
689.
Chamberland, M. E., Lyons, R. W., and Brock, S. M. (1977).
Effect of inline filtration of intravenous infusions on the inci-
dence of thrombophlebitis. Am. J. Hosp. Pharm., 34, 1068.
Cheney, F. W. jr, and Lincoln, J. R. (1964). Phlebitis from plas-
tic intravenous catheters. Anesthesiology, 25, 650.
Clark, P. B., Polak, A., and Hajnal, J. A. L. (1960). A clinical use
of hydrocortisone in the prevention of postinfusion throm-
bophlebitis. Postgrad. Med.J., 36, 55.
Collin, J., Collin, C , Constable, F. L., and Johnston, I.D.A.
(1975). Infusion thrombophlebitis and infection with various
cannulas. Lancet, 2, 150.
Tweedle, D. E. F., Venables, C. W., Constable, F. L., and
Johnston, I. D. A. (1973). Effect ofamillipore filter on compli-
cations of intravenous infusions; a prospective clinical trial. Br.
Med.J., 4,455.
Crunkhorn, P., and Willis, A. L. (1971). Cutaneous reactions to
intradermal prostaglandins. Br.J. Pharmacol., 41, 49.
Curry, J. T., and Zallen, R. D. (1973). Reduction of throm-
bophlebitis associated with indwelling catheters. J. OralSurg.,
31,636.
Daniell, H. W. (1973). Heparin in the prevention of infusion
phlebitis.7.4.M./l.,226, 1317.
Davis, N. M., and Turco, S. J. (1973). Paniculate matter in
intravenous fluids—phase three. Am. J. Hosp. Pharm., 30,
231
611.
De Luca, P. P., Rapp, R. P., Bivins, B., McKean, H. E., and
Griffen, W. O. (1975). Filtration and infusioin phlebitis: a
double-blind prospective clinical study. Atn.J. Hosp. Pharm.,
32, 1001.
Denis, D., Charleson, S., Rackham, A., Jones, T. R., Ford-
Hutchinson, A. W., Lord, A., Cirino, A., Girard, Y., Larve,
M., and Rokach, J. (1982). Synthesis and biological activities
of leukotriene F4 and leukotriene F4 sulfone. Prostaglandins,
24,801.
Dinley, R. J. (1976). Venous reactions related to in-dwelling plas-
tic cannulae: a prospective clinical trial. Curr. Med. Res. Opin.,
3, 607.
Eerola, R., and Pontinen, P. J. (1964). Prophylaxis of infusion
thrombophlebitis. Ann. Chir. Gynaec. Fenn., 53, 519.
Elfving, G., and Saikku, K. (1966). Effect of pH on the incidence
of infusion thrombophlebitis. Lancet, 1, 953.
Eremin, O., and Marshall, V. (1977). Complications of intraven-
Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017
ous therapy: reduction by buffering of intravenous fluid prep-
aration. Med. J. Aust., 2, 528.
Erwin, P., Strickler, J. A., and Rice, C. O. (1953). Use of
polyethylene tubing in intravenous therapy for surgical
patients. Arch. Surg., 66, 673.
Evans, W. E., Barker, L. F., and Simone, J. V. (1976). Double-
blind evaluation of 5 jxm final filtration to reduce postinfusion
phlebitis. Am. J. Hosp. Pharm., 33, 1160.
Ferguson, R. L., Rosett, W., Hodges, G. R., and Barnes, W. G.
(1976). Complications with heparin-lock needles. A prospec-
tive evaluation. Ann. Intern. Med., 85, 583.
Fonkalsrud, E. W., Carpenter, K., Masuda, J. Y., and Becker-
man, J. H. (1971). Prophylaxis against postinfusion phlebitis.
Surg. Gynecol. Obstet., 133, 253.
Murphy, J., and Smith, F. G. jr. (1968). Reduction of infu-
sion thrombophlebitis with buffered glucose solutions.
Surgery, 63, 280.
Frazer, I. H., Eke, N., and Laing, M. S. (1977). Is infusion
phlebitis preventable? Br. Med.J., 2, 232.
Garvan, J. M., and Gunner, B. W. (1963). Intravenous fluids: a
solution containing such particles must not be used. Med. J.
Aust., 2, 140.
Ghildyal, S. K., Pande, R. C , and Misra, T . R. (1975). His-
topathology and bacteriology of postinfusion phlebitis. Intern.
Surg., 60, 341.
Gjares, J. E. (1957). Narkotal thrombophlebitis. Svenska, Lak-
tidn., 50, 3825.
Gritsch, H. J., and Ballinger, C. M. (1959). Value of indwelling
catheters in intravenous therapy. J.A.M.A., 171, 281.
Gryglewski, R. J., Bunting, S., Moncada, R. J., and Vane, J. R.
(1976). Arterial walls are protected against deposition of
platelet thrombi by a substance (prostaglandin X). Prostaglan-
dins, 12, 685.
Handfield-Jones, R. P. C , and Lewis, H. B. M. (1952). Rubber
tubing as cause of infusion thrombophlebitis. Lancet, 1, 585.
Hastbacka, J., Tammisto, T., Elfving, G., and Tiitinen, P.
(1965). Infusion thrombophlebitis, a clinical study based on
1048 cases. Ada Anaesthesiol. Scand., 10, 9.
Hecker, J. F. (1980). Thrombus formation on cannulae. Anaesth.
Intens. Care, 8, 187.
Edwards, R. O. (1981). Effects of roughness on the throm-
bogenicity of a plastic. J. Biomed. Mat. Res., 15, 1.
Lewis, G. B. H., and Stanley, H. (1983). Nitroglycerine
ointment as an aid to venepuncture. Lancet, 1, 332.
Hegarty, J. E., and Dundee, J. W. (1977). Sequelae after the
intravenous injection of three benzodiazepines—diazepam,
lorazepam and flunitrazepam. Br. Med.J., 2, 1384.
232 BRITISH JOURNAL OF ANAESTHESIA

Hessov, I., Allen, J., Arendt, K., and Gravholt, L. (1977). Infu-
sion thrombophlebitis in a surgical department. Acta Chir.
Scand., 143, 151.
Bojsen-Moller, M. (1976). Experimental infusion throm-
bophlebitis. Importance of the pH of glucose solutions.
Eur.J. Intern. CareMed., 2, 97.
Hewitt, J. C , Hamilton, R. C , O'Donnell, J. F., and Dundee,
J. W. (1966). Clinical studies of induction agents XIV: A com-
parative study of venous complications following thiopentone,
methohexitone and propanidid. Br. J. Anaeslh., 38, 115.
Hohn, A. R. (1966). Continuous venous catheterization in chil-
dren. J.AM.A., 197, 658.
Horvitz, A., Sachar, L. A., and Elman, R. (1943). An experi-
mental study of phlebitis following venoclysis with glucose and
amino acid solutions. J . Lab. Clin. Med., 28, 842.
Inagaki, J., and Bodey, G. P. (1973). Phlebitis associated with
cephalosporins: cephapirin versus cephalothin. Curr. Ther.
Res., 15, 37.
Mason, D. T., and Braunwald, E. (1965). The effects of nitro-
glycerine and amyl nitrate on arteriolar and venous tone in the
human forearm. Circulation, 32, 755.
Mattila, M. A. K., Rossi, M. L. and Ruoppi, M. K. (1981).
Reduction of venous sequelae of i.v. diazepam with a fat emul-
sion as solvent. Br.J. Anaesth., 53, 1265.
Medical Research Council's Subcommittee (1957). Thrombo-
phlebitis following intravenous infusions: A trial of plastic
and rubber giving sets. Lancet, 1, 595.
Michaels, L., and Ruebner, B. (1953). Growth of bacteria in
intravenous infusion fluids. Lancet, 264, 772.
Mikkelsen, H., Hoel, T. M., Bryne, H., and Krohn, C. D.
(1980). Local reactions after i.v. injectios of diazepam, flunit-
razepam and isotonic saline. Br.J. Anaesth., 52, 817.
Moncada, S., Gryglewski, R., Bunting, S., and Vane, J. R.
(1976). An enzyme isolated from arteries transforms prosta-
glandin endoperoxides to an unstable substance that inhibits
platelet aggregation. Nature (Land.), 263, 661.

Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017

James, J. D. (1954). Hazards of blood transfusion. Lancet, 2,767. Moncada, S.,and Vane, J. R. (1977). The discovery of prostacyc-
Jamieson, D., Desjardins, C , and Caron, J. (1972). Thrombo- lin (PGX): A fresh insight into arachidonic acid metabolism; in
phlebitis following thiopentone injection. Can. Anaesth. Biochemical Aspects of Prostaglandins and Thromboxanes (eds J.
Soc.J., 19,659. Fried and N. Karasch), p. 155. Santa Monica: Intersci Prostag-
Jones, P. F. (1954). Thrombophlebitis following intravenous landin Symp.
infusions. Lancet, 2, 970. (1978). Prostacyclin, platelet aggregation and throm-
(1957). Thrombophlebitis following intravenous infusion— bosis; in Platelets: A Multidisciplinary Approach (eds G. de
a review of its aetiology and prevention. Postgrad. Med.J., 33, Gaetanoand S. Garentini), p.239. New York: Raven Press.
444. Nordell, K., Mogensen, L., and Nyquist, O. (1972). Throm-
Kakkar, V. V., Howe, C. T., Flanc, C , and Clarke, M. B. bophlebitis following intravenous lignocaine infusion. Acta
(1969). Natural history of post-operative deep-vein throm- Med. Scand., 192,263.
bosis. Lancet, 2, 230. Norden, C. W. (1969). Application of antibiotic ointment to site
Kay, I., and Roberts, S. S. (1967). Infusions and phlebitis. Am. of venous catheterization: A controlled trial. J. Infect. Dis.,
J. Nursing, 67, 2081. 120,611.
Kelton, J. G., Hirsh, J., Carter, C. J., and Buchanan, M. R. O'Donnell, J. F., Hewitt, J. C , and Dundee, J. W. (1969). Clin-
(1978). Thrombogenic effect of high-dose aspirin in rabbits. J . ical studies of induction agents XXVIII: A further comparison
Clin. Invest., 62,892. of venous complications following thiopentone,
Kenny, G. N. C. (1981). Alternatives to inhalation anaesthesia. methohexitone and propanidid. Br. J. Anaeslh., 41, 681.
Br.J. Anaeslh., 53, 75S. Page, B. H., Raine, G., and Jones, P. F. (1952). Thrombo-
Kivalo, I., and Tammisto, T. (1959). Use of pressuren in Caesa- phlebitis following intravenous infusions. Lancet, 2, 778.
rean section. Acta Anaesthesiol. Scand., 2, 35. Pederson, B. M. (1970). A solution for post-infusion thrombo-
Lane, A. Z., Taggart, J. G., and lies, R. L. (1972). Relative inci- phlebitis. Am.J. Nursing, 70, 325.
dence of phlebitis caused by continuous intravenous infusions Perkins, R. L., and Saslaw, A. (1966). Experiences with
of cephapirin and cephalothin. Antimicrob. Agents Chemother., cephalothin. Ann. Intern. Med., 64, 13.
2,234. Polok, A. (1956). Hydrocortisone in the prevention of transfu-
Langdon, D. E., Harlan, J. R., and Bailey, R. L. (1973). Throm- sion thrombophlebitis. Lancet, 1, 484.
bophlebitis with diazepam used intravenously. J.A.M.A., Pussell, B. A., and Pitney, W. R. (1972). The use of anticoagul-
223, 134. ants in the prevention of venous thrombosis after myocardial
Lees, N. W., Glasser, J., McGroarty, F. J., and Miller, B. M. infarction. Med.J. Aust., 2, 1101.
(1981). Etomidate and fentanyl for maintenance of anaesthesia. Ravitch, M. M. (1969). Total intravenous alimentation. Med.
Br.J. A naesth., 53,959. Times, 97, 238.
Levin, R. I., Jaffe, E. A., Welzsler, B. B., and Jack-Goldman, K. Ross, S. A. (1972). Infusion phlebitis. Nursing Res., 21, 313.
(1981). Nitroglycerine stimulates synthesis of prostacyclin by Roth, G. J., and Majerus, P. W. (1975). The mechanism ot the
cultured human endothelial cells. J. Clin. Invest., 67, 763. effect of aspirin on human platelets I. Acetylation of a particu-
Lewis, G. B. H., and Hecker, J. F. (1984). Changes in local ven- late fraction protein, j . Clin. Invest., 54, 624.
ous tone in response to infusions of saline and dextrose infu- Roy, R. B., Wilkinson, R. H., and Bayliss, C E. (1967). Utiliza-
sions. Anaeslh. Inlens. Care, 12, 27. tion of long nylon catheters for prolonged intravenous infu-
Little, J. M., Loewenthal, J., and Mansfield, A. O. (1974). sions. Can. Med. Assoc.J., 96,94.
Thromboembolic conditions of the venous system; in Scientific Ryan, P. B., Rapp, R. P., and De Luca, P. P. (1973). Inline final
Foundations of Surgery, 2nd Edn, p.419. London: Heinemann. filtration—a method of minimizing the risk of bacterial, fungal
McNair, T. J., and Dudley, H. A. F. (1959). The local complica- and paniculate contamination in intravenous therapy. Bull.
tions of intravenous therapy. Lancet, 2, 365. Parenter. Drug. Assoc, 27, 1.
Martin, J. T. (1965). Plastic devices for intravascular therapy. Samuelsson, B. (1980). The leukotrienes: a new group of biolog-
Anesth. Analg., 44, 25. ically active compounds including SRS-A. Tips Reviews,
Martin, P. G. C. (1944). Heparin in intravenous infusions, (Dec), 227.
including penicillin therapy. Br. Med.J., 2, 308. Schafermeyer, R. W. (1974). Prevention of phlebitis. J.A.M.A.,
INFUSION THROMBOPHLEBITIS 233

228,695. phlebitis. Aneslh. Analg., 49, 150.

Sharpey-Scafer, E. P., and Ginsburg, J. (1962). Humoral agents Tse, R. L., and Lee, M. W. (1971). pH of infusion fluids: a pre-
and venous tone: Effects of catecholamines, 5HT, histamine disposing factor in thrombophlebitis. J.A.M.A., 215, 642.
and nitrites. Lancet, 2, 1337. Turco, S. J., and Davis, N . M. (1971). A study of paniculate
Siebert, W. T., Westerman, E. L., Smilack, j . D., Bradshaw, matter in intravenous infusion fluids—phase two. Am. J.
M. W.,and Williams, T . W . (1976). Comparison of thrombo- Hosp. Pharm., 28, 620.
phlebitis associated with three cephalosporin antibiotics. Vere, D. W., Sykes, C. H . , and Armitage, P. (1960). Venous
Antimicrob. Agents Chem., 10, 467. thrombosis during dextrose infusion. Lancet, 2,627.
Siebke, H., Ellertsen, B. B., and Lind, B. (1976). Reactions to Walters, M. B., Stanger, H. A. D., and Rotem, C. E. (1972).
intravenous injections of diazepam. Br.J. Anaesth., 48, 1187. Complications with percutaneous central venous catheters.
Skajaa, T., Dahl, J., Jensen, J. K., and Kvisselgaard, N. (1961). J.A.M.A., 220, 1455.
(Cited by Hastbacka, J. et al. (1965) in: Infusion Throm- Warthen, H. J. (1930). Int. Surg. Dig., 10, 3. (quoted by P. F.
bophlebitis. Ada Anaesthesiol. Scand., 10, 9.) Jones, 1957).
Sketch, M. H., Cale, M., Mohiuddin, S. M., and Booth, R. W. Watt, A. G. (1977). A critical examination of the insertion of
(1972). Use of percutaneously inserted venous catheters in intravenous cannulae. Med.J. Aust., 1, 111.
coronary care units. Chest, 62, 684. Wessler, S., and Rogers, W. R. (1956). Intermittent intravenous
Sodeman, W. A.,and Sodeman, W. A. jr(Eds)(1967).Pathologic therapy: Experience with a simple method for the administra-
Physiology. Mechanisms of Disease. London: W. B. Saunders. tion of heparin and other drugs in 269 medical and surgical

Downloaded from http://bja.oxfordjournals.org/ by guest on January 12, 2017

Spanos, H. G., and Hecker, J. F. (1976). Thrombus formation patients. N. Engl. J. Med., 255, 22.
on indwelling venous cannulae in sheep: effects ot time, size Williams, J. T., and Moravec, D. F. (1967). Intravenous Therapy.
and materials. Anaeslh. Intens. Care, 4, 217. Chicago: Clissold Publishing Co.
Stein, J. M., and Pruitt, B. A. (1970). Suppurative thrombo- Wilmore, D. W., and Dudrick, S. J. (1969). Safe long term ven-
phlebitis. A lethal iatrogenic disease. N. Engl. J. Med., 282, ous catheterization. Arch. Surg., 98, 256.
1452. Woodhouse, C. R. J. (1979). Movelat in the prophylaxis of infu-
Stephen, M., Loewenthal, J., Wong, J., and Benn, R. (1976). sion thrombophlebitis. Br. Med. J., 2,454.
Complications of intravenous therapy. Med.J. Aust., 2, 557. (1980). Infusion thrombophlebitis: the histological and clin-
Storey, G. (1971). Clinical manifestations of chancroid. Br. J. ical features. Ann. R. Coll. Surg. Engl., 62, 364.
Urol., 42, 738. Wyant, G. M., and Studney, L. J. (1970). A study of diazepam
Swanson, J. T., and Aldrete, J. A. (1969). Thrombophlebitis (Valium) for induction of anaesthesia. Can. Anaesth. Soc. J.,
after intravenous infusion—Factors affecting its incidence. 17. 166.
Rocky Mountain Med. J., 66,48. Zinner, S. H., Denny-Brown, B. C , Braun, P., Burke, J. P.,
Thayssen, P. (1973). Postinfusion phlebitis and the calibre of the Toala, P., and Kass, E. H. (1969). Risk of infection with
catheter. UgeskrLaeg., 135, 1238. intravenous indwelling catheters: effects of application of
Thomas, E. T., Evers, W., and Racz, M. B. (1970). Postinfusion antibiotic ointment. J. Infect. Dis., 120, 616.