Idea Transcript
lAC
Journal of Antimicrobial Chemotherapy (1997) 39, 371-381
Risk factors for adverse cutaneous reactions associated with intravenous vancomycin Tony M. Korman*, John D. Turnidge and M. Lindsay Grayson Department of Infectious Diseases and Microbiology, Monash Medical Centre, Clayton, VIC 3168, Australia
Introduction Over the past decade, f1-lactam-resistant Gram-positive organisms have emerged as significant nosocomial pathogens and vancomycin usage has increased considerably. The widespread use of vancomycin may result in increased morbidity from adverse drug reactions. 1 Adverse cutaneous reactions are the most common reason for the cessation of vancomycin therapy.2 Infusionrelated 'red man syndrome' occurs in 80-90% of healthy volunteers 3- 5 but rates in patients vary from 0% to 50%.6--8 The incidence of non-infusion-related cutaneous reactions is not well established. Patient factors predisposing for adverse cutaneous reactions have not been identified in clinical studies. Patients with reactions to one brand of vancomycin may tolerate other brands without complications; however, few studies have compared the incidence of reactions with different brands. During early 1994 an apparent increase in the incidence of vancomycin-associated cutaneous reactions was observed at our institution. Subsequently, we discovered that there had been a recent change in the brand of vancomycin used. This prompted us to undertake a retrospective study to investigate whether the incidence of reactions had actually increased after the change in the brand of vancomycin.
Other objectives were to characterize the clinical features and identify potential risk factors for adverse cutaneous reactions associated with vancomycin therapy.
Patients and methods Patient selection Monash Medical Centre is a 747-bed acute community and tertiary referral centre in Melbourne, Australia. Patients over 16 years of age who had received iv vancomycin during a 14-month period from June 1993 to July 1994 were identified by review of in-patient pharmacy medication profiles. Patients who had received vancomycin in the intensive care unit or operating theatre were excluded from the analysis because adverse cutaneous reactions, in particular infusion-related reactions, may be difficult to recognize in these patients. 6
Definitions The duration of a course was defined as the number of days between the administration of the first and last doses of vancomycin (inclusive). When not administered daily (e.g. in patients with renal impairment), the duration
*Tel: +61-3-9550-4564; Fax: +61-3-9550-4533.
371
Downloaded from https://academic.oup.com/jac/article-abstract/39/3/371/668331 by guest on 14 April 2019
We retrospectively studied adverse cutaneous reactions associated with intravenous vancomycin therapy over a 14-month period when two different brands of vancomycin were used. Of 224 adults, 12 (5.4%) had infusion-related reactions; ten of 174 patients who received more than one day of vancomycin (5.7%) had delayed cutaneous reactions. Age less than 40 years was a risk factor for both infusion-related and delayed reactions by both univariate and multivariate analysis. Duration of therapy greater than 7 days was a risk factor for delayed reactions. There was a significant increase in adverse cutaneous reactions associated with the use of a particular batch of vancomycin, although analytical testing of this batch failed to identify any difference from other batches associated with routine rates of adverse reactions. Awareness of vancomycin-associated infusion-related and delayed cutaneous reactions is necessary, and the risk factors associated with these reactions may have important clinical implications.
T. M. Korman et al.
Patients and courses The clinical records of patients receiving a course of vancomycin were reviewed and the following data were obtained: age, sex, weight, underlying disease(s), concomitant drug therapy, previous history of drug allergy and duration of therapy. The initial dose of vancomycin administered was calculated as mg/kg/dose, and the subsequent dose calculated as mg/kg/day. The actual duration of vancomycin infusions was not routinely documented in patient records; however, hospital guidelines dictate that a 1 g dose of vancomycin is given as an infusion in at least 500 mL of fluid over 1 h.
Adverse drug reactions Suspected adverse drug reaction reports generated by nursing, pharmacy or medical staff caring for the patients
were examined. In addition, patient charts were reviewed for any adverse reactions documented during vancomycin courses. The following details of the suspected adverse cutaneous reactions were collected: day of onset during vancomycin therapy, clinical features (including description of skin rash, pruritus, fever attributed to vancomycin), associated eosinophilia (increase in the absolute eosinophil count to > 0.44 X 109/L), the treatment required (if any) and the outcome of the reaction. Details of any attempts at rechallenge and further management were recorded.
Serum vancomycin concentrations Serum vancomycin concentrations determined by enzyme multiplied immunoassay technique (EMIT) were recorded. The mean peak and trough levels measured during each course were calculated. If the trough level measured less than the sensitivity of the assay «5 mg/L), an arbitrary level of 4 mg/L was used for calculations. If the peak level measured >80 mg/L an arbitrary level of 81 mg/L was used for calculations.
Vancomycin brands and analytical testing of batches Pharmacy records were used to ascertain the brand and batch of vancomycin used for each course. The brands of vancomycin used included: Vancocin CP (Eli Lilly, West Ryde, New South Wales, Australia), Vancoled (Lederle Laboratories, Baulkham Hills, New South Wales, Australia) and vancomycin hydrochloride (Faulding Pharmaceuticals/David Bull Laboratories, Mulgrave North, Victoria, Australia), hitherto referred to as DEL vancomycin. Product specifications for minimum vancomycin B content were 95% for the chromatographically purified Vancocin CP, and 88% for both Vancoled and DBL vancomycin. Analytical testing of DBL vancomycin batches was performed by the manufacturer in response to adverse drug reaction reports forwarded to the company by our institution. The vancomycin B content and impurity profiles of batches of raw material and finished product were analysed by HPLC.
Statistical analysis Data were evaluated by Student's t-test, X2 test or Fisher's exact test where appropriate. P < 0.05 was considered significant. Significant univariate risk factors for adverse reactions were analysed by stepwise multiple logistic regression using a computed model (TRUE EPIST AT Version 5, Epistat services, Richardson, TX, USA, 1994). Data are presented as mean:±: standard deviation.
372
Downloaded from https://academic.oup.com/jac/article-abstract/39/3/371/668331 by guest on 14 April 2019
included days when detectable serum vancomycin levels would have been maintained. Adverse cutaneous reactions were defined as either infusion-related or delayed. An infusion-related reaction was defined as the development of symptoms or signs (e.g. pruritus, erythema) during iv infusion of vancomycin. A delayed reaction was defined as the development of a skin rash during a course of vancomycin therapy, not temporally related to iv infusion. Only patients who received more than one day of vancomycin were included in the analysis of delayed reactions. Since thrombophlebitis would be difficult to identify accurately in a retrospective study, it was not recorded. In order to assess the likelihood that the adverse reactions were caused by vancomycin, each was allocated a causality rating using the following standard published criteria: 9 (i) Certain: confirmed by rechallenge; and/or confirmed by laboratory data; and/or reaction onset immediately followed drug administration (within 5 min if injection was the method of administration); (ii) Probable: temporal or spatial (e.g. skin) correlation with administration; and/or recovery on withdrawal of drug if no other drug is withdrawn and no therapy given; and/or uncommon clinical phenomenon associated with the administration of drug in the absence of other factors; (iii) Possible: a possible alternative explanation exists; and/or more than one drug is suspected; and/or data are incomplete and/or recovery follows withdrawal of more than one drug; and/or time relationship is not clear; and/or outcome of the reaction is not recorded; and/or recovery follows therapy in addition to withdrawal of drug; (iv) Causality unclear: this classification is accorded where a clinical event may well be explained as arising from factors related to underlying disease, or other non-drug aetiology. Only suspected adverse reactions allocated the rating of certain, probable or possible were included in the analysis of vancomycin-associated adverse reactions.
Vancomycin-associated cutaneous reactions
Results Patients and courses During the study period 269 patients commenced a course of vancomycin, of whom the records of 267 patients (99.3 %) were available for analysis. Of these 267 patients, 43 who had received vancomycin in the intensive care unit (n = 41) or the operating room (n = 2) were excluded. For the 224 patients analysed, mean age was 56.0 ± 19.0 years (range 16-90; 110 males and 114 females), and duration of vancomycin therapy was 7.5 ± 9.3 days.
further infusion~related adverse reactions, but developed an erythematous maculopapular rash, pruritus, fever and eosinophilia on day 23 of vancomycin therapy. The reaction resolved after stopping vancomycin (along with ceftriaxone and doxycycline) after 42 days. Patient 12 had no further infusion-related adverse reactions but developed a widespread bullous eruption, fever and eosinophilia on day 28 of vancomycin therapy. The eruption subsided after ceasing vancomycin, and a course of ftucloxacillin was tolerated without complications.
Twelve of the 224 patients (5.4%) had infusion-related reactions; ten patients out of 174 (5.7%) who received more than 1 day of vancomycin had delayed cutaneous reactions, including two patients who had both infusionrelated and delayed reactions. Thus, a total of 20 of the 224 patients (8.9%) had a vancomycin-associated adverse cutaneous reaction. Details of the patients with vancomycin-associated adverse cutaneous reactions are shown in Table 1.
Infusion-related cutaneous reactions Thirteen patients with infusion-related adverse reactions were identified. One reaction was excluded as the patient developed pruritus and erythema when her third dose of vancomycin was inadvertently administered over 10 min, but she tolerated previous and subsequent doses administered over at least 60 min. Therefore, 12 of the 224 (5.4 % ) patients were defined as having infusion-related reactions. All 12 patients complained of pruritus, and ten had accompanying skin erythema. Patient 5 had a past history of a vancomycin infusion-related reaction and developed a nonpruritic erythematous maculopapular skin rash which resolved 2 days after a single vancomycin infusion. In the other 11 patients, the signs and symptoms abated soon after ceasing the infusion. Four patients were treated with Hrreceptor antagonists. By definition, all the infusionrelated reactions were accorded the causality rating 'certain'. Four patients in this group received no further doses of vancomycin. Five patients tolerated continued vancomycin therapy at infusion rates of 500-1000 mg/h without further reactions. The other three patients developed adverse cutaneous reactions after receiving further vancomycin. Patient 2 had an infusion-related reaction associated with DBL vancomycin, was rechallenged with Vancoled, and developed pruritus and erythema after 5 mins of vancomycin infusion (rate of 500 mg/h). She tolerated an 8-week course of iv teicoplanin without an adverse cutaneous reaction. Two patients who continued prolonged courses with DBL vancomycin despite having an infusion-related adverse reaction developed delayed reactions. Patient 11 tolerated vancomycin without
Thirteen patients with delayed cutaneous reactions were identified. Three reactions related to underlying disease or other non-drug aetiology, i.e. 'causality unclear', were excluded from further analysis. Therefore, ten patients had delayed reactions or 5.7% of the 174 who received vancomycin for more than 1 day. Nine of 74 patients (12.2%) who received vancomycin for >7 days and eight of 32 patients (25%) treated for ;;.14 days experienced delayed reactions. The mean onset of the delayed reactions was 20.1 ± 8.8 days (range: 6-31 days) after the commencement of vancomycin therapy. Cutaneous manifestations of the delayed reactions included: erythematous skin rashes (n = 8), exfoliative dermatitis (n = 1), bullous eruption (n = 1), facial oedema (n = 1) and pruritus (n = 5). Fever attributed to vancomycin was noted in eight patients, one of whom also had rigors, myalgias and a dry cough. One patient developed acute renal failure concurrently. Six of ten patients (60%) had associated eosinophilia. Delayed reactions resolved in seven patients after ceasing vancomycin alone. Two of these reactions were confirmed on rechallenge with vancomycin (see below) and hence they were allocated the causality rating 'certain'. The other five reactions resolved after the withdrawal of vancomycin alone, but were not confirmed by rechallenge and hence were allocated the causality rating 'probable'. In three patients, the adverse reaction was considered 'possible' as a resolution followed the cessation of vancomycin and at least one other drug. Four patients received specific therapy for the adverse reactions: two received Hrreceptor antagonists alone and two required Hrreceptor antagonists and treatment with systemic corticosteroids for 1-3 days. Three patients with delayed adverse cutaneous reactions associated with DBL vancomycin were rechallenged with vancomycin. Two patients had recrudescence of pruritic, erythematous skin rashes promptly after rechallenge with DBL vancomycin (patient 19) or Vancocin CP (patient 15). The third patient (patient 20) had a reaction with DBL vancomycin but tolerated a 42-day course of Vancoled without a skin rash before developing eosinophilia which resolved only after cessation of Vancoled.
373
Downloaded from https://academic.oup.com/jac/article-abstract/39/3/371/668331 by guest on 14 April 2019
Delayed cutaneous reactions Adverse cutaneous reactions
~ Pen
no
acute myeloid leukaemia none
F
63
33
5
6
none
reflux nephropathy thalassaemia intermedia
M
F
60
20
33
9
10
11
F
none
F
16
8
F
26
7
no
no
no
no
Van
no
Crohn's disease gastric carcinoma
M
67
4
M
no
none
M
20
3
no
none
F
23
Pen
none
F
2
Past drug allergy
Underlying Sex disease(s)
20
Age (years)
1
Patient no.
ceftriaxone, doxycycline, ranitidine, warfarin
ampicillin, warfarin, gentamicin none
gentamicin, phenytoin, ranitidine none
medroxyprogesterone, oestrogen allopurinol
hydrocortisone
none
metronidazole, gentamicin warfarin
Concomitant medications
DBL (324) DBL (324)
DBL (324)
DBL (324)
DBL (324)
IR
DBL (324) DBL (324) VLD
DBL (324)
DR
IR
IR
IR
IR
IR
IR
IR
IR
IR
IR
Type of reaction
VLD
pyelonephritis DBL (324) endocarditis DBL (324)
septic arthritis prophylaxis
prophylaxis
prophylaxis
VPshunt infection VC-related septicaemia VC-related septicaemia
endocarditis
endometritis
Indication for Brand vancomycin (batch)
23
1
1
1
1
1
1
1
1
1
1
1
Day of onset
rash, pruritis, fever, eosinophilia
erythema, pruritis pruritis
erythema, pruritis erythema, pruritis
erythema
erythema, pruritis
erythema, pruritis erythema, pruritis erythema, pruritis erythema, pruritis erythema, rash
Features of reaction
no yes (no IR or DR) no
no
yes (no IR)
yes (no IR or DR) yes (no IR)
no yes (no IR, but later DR)
no
Vanc none Vanc,d
Vanc
none
none
Vanc none
none
Van
certain
yes (IR)b
Van
possible
certain
certain
certain
certain
certain
certain
certain
certain
certain
certain
Causality rating"
yes (IR)
Rechallenge (outcome)
none
Medications ceased
Table I. Patient details and features of vancomycin-associated cutaneous adverse reactions
Downloaded from https://academic.oup.com/jac/article-abstract/39/3/371/668331 by guest on 14 April 2019
:--0
~
~
! ....
~
..=a
~
.3:
tDI!
-.JI
("j.)
M
54
75
22
26
13
14
15
16
F
70
42
30
18
19
20
no
no
none
chronic renal failure diabetes mellitus none
none
none
no
no
no
Pen
no
haemono dialysis epidermolysis no buHosa
hypersensitivity pneumonitis
thalassaemia major
ticarcillinl davulanate, frusemide, ranitidine imipeneml cilastin, dothiepin ranitidine imipeneml cilastin
none
ceftriaxone
none
bumetanide
hydrocortisone, digoxin
ceftriaxone, desferrioxamine
osteomyelitis
osteomyelitis
osteomyelitis
endocarditis
osteomyelitis
VC-related septicaemia VC-related septicaemia
osteomyelitis
osteomyelitis
DR
DR
DR
DR
DR
DR
DR
DR
. _ - - _...
(324)
DBL
(324)
DBL
VLD
(324)
DBL
(324)
DBL
(324)
DBL
VLD
(324)
DBL
10
13
25
29
23
13
6
31
rash, fever
rash, pruritis, fever
probable
possible
no
no
orDRe )
yes (no IR
certain yes OR)
Van
probable
no
probable
certain
possible
no
yes OR)
probable
no
yes (no IR, certain but later DR) no probable
Van d
bullous none eruption, fever, eosinophilia Van rash, fever, acute renal failure exfoliative Van, bumetanide dermatitis rash, Van fever, eosinophilia Vanc,d rash, pruritis, fever, rigors, myalgia, cough eosinophilia Vanc,d rash, pruritis, facial, oedema, eosinophilia rash, Van, pruritis, ticarcillinl eosinophilia davulanate C
28
DR
VLD
none
pruritis
1
IR
VLD
DEL, DEL vancomycin; DR, delayed cutaneous reaction; JR, infusion-related reaction; Pen, penicillin; Van, vancomycin; VC, vascular catheter; VLD, Vancoled; VP, ventriculo-peritoneal. a See 'Patients and methods' for definitions (infusion-related reactions are classified as 'certain'). b Tolerated 8 weeks of teicoplanin without infusion-related reactions. c Patient received promethazine treatment. d Patient received corticosteroid (oral prednisolone or iv hydrocortisone) treatment. e Developed eosinophilia which resolved after cessation of vancomycin.
F
F
17
17
M
F
F
F
M
25
12
Downloaded from https://academic.oup.com/jac/article-abstract/39/3/371/668331 by guest on 14 April 2019
~
1!r :I '"
\"l
~ ~
-
~
""
~
~
= t:J
~
=:....
i"l
Q",
"* ~
~o
OJ)
.S u 0:
u