Annals of Medicine and Surgery 49 (2020) 37–40

Contents lists available at ScienceDirect

Annals of Medicine and Surgery

journal homepage: www.elsevier.com/locate/amsu

Cellulitis in older people over 75 years – are there differences? T

a,∗ a,b a a
Manoj Kumar , Vincent Jiu Jong Ngian , Clarence Yeong , Caitlin Keighley ,
Huong Van Nguyena,b, Bin Soo Onga,b
a
Bankstown-Lidcombe Hospital, Bankstown, New South Wales, Sydney, Australia
b
University of New South Wales, Sydney, New South Wales, Australia

ARTICLE INFO ABSTRACT

Keywords: Aim: To examine differences in risk factors, clinical features and outcomes of cellulitis between those 75 + years
Cellulitis and those < 75 years admitted to a metropolitan hospital.
Older people Methods: A prospective study of patients with limb cellulitis requiring intravenous antibiotics conducted at
Outcomes Bankstown-Lidcombe Hospital, Australia from June 2014 to April 2015.
Results: Thirty one patients were 75 + years and 69 less than 75 years. A greater proportion of older patients
resided in nursing home (25.8% vs 2.9% respectively, p = 0.001) and mobilized with walking aid(s) (58.1% vs
11.6% respectively, p < 0.001). Significantly more older patients had documented hypertension (45.2% vs
23.2% respectively p = 0.035), atrial fibrillation (33.5% vs 5.8% respectively, p < 0.001), dementia (22.6% vs
1.4% respectively, p = 0.001) and malignancy (16.1% vs 1.4% respectively, p = 0.010). The clinical pre-
sentation of cellulitis and cellulitis severity (Eron classification) did not significantly differ in both groups;
however older patients were more likely to have dependent edema (OR 4.0, 95%CI 1.3–12.6, p = 0.018) and
less likely to be obese (OR 0.3, 95%CI 0.1–0.8, p = 0.012) or had a past history of cellulitis (OR 0.3, 95%CI
0.1–1.0, p = 0.044) on presentation. Despite the age difference, there were no major differences in intravenous
antibiotic choice, hospital length of stay, and hospital readmission rates in both groups. Older patients however,
were more likely to experience complications such as falls and/or decreased mobility (38.7% vs 15.9% re-
spectively, p = 0.020) during the cellulitis episode.
Conclusion: There are minor differences in the risk factors and clinical features of cellulitis in older patients as
compared to the young. Outcomes are similar except for a higher incidence of hospital related complications.

1. Introduction (including use of antibiotics); however age-related pharmacokinetics

and pharmacodynamics, cognitive status and social circumstances [11]
Cellulitis is a bacterial infection of the skin involving the dermis and may impact on treatment decisions particularly need for hospitaliza-
subcutaneous fat. In Australia, cellulitis accounts for over 250,000 tion.
hospital bed days, or 10.5% of potentially preventable hospitalizations Once hospitalized, age is an independent risk factor for increased
[1].While most episodes of cellulitis can be managed as an outpatient, a length of stay for cellulitis with other factors being long duration of
significant proportion, particularly older people, require hospitaliza- symptoms, tachycardia, hypotension, leukocytosis, hypoalbuminemia,
tion. Over a 12-month period from 2014 to 2015, the cellulitis hospi- elevated serum creatinine, bacteremia, obesity and diabetes mellitus
talization rate was 1100 per 100,000 in the 80 plus age group as op- [12–15].
posed to 237 episodes per 100,000 in the general population [1]. Age is significantly associated with increased mortality from cellu-
Cellulitis typically presents with pain, erythema, warmth and litis although it is unclear if this is due to illness severity or underlying
edema. Systemic symptoms including fever and tachycardia may be comorbidity [16]. Other factors associated with mortality are delayed
present although thought to be less frequent in older persons [2–6]. administration of antibiotics, presence of multiple comorbidities, pre-
Known risk factors for cellulitis are venous edema, lymphedema, skin vious myocardial infarction, congestive heart failure, liver disease,
conditions, traumatic injury, leg ulcers, peripheral vascular disease, hypoalbuminemia, renal insufficiency, morbid obesity, lower limb
fungal infections, past history of cellulitis and obesity [7–10]. edema, Pseudomonas aeruginosa infection, bacteremia and septic shock
Age alone does not alter treatment principles for bacterial cellulitis [14,17].

∗
Corresponding author.
E-mail address: [email protected] (M. Kumar).

https://doi.org/10.1016/j.amsu.2019.11.012
Received 12 September 2019; Received in revised form 13 November 2019; Accepted 17 November 2019
2049-0801/ © 2019 The Authors. Published by Elsevier Ltd on behalf of IJS Publishing Group Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/BY/4.0/).
M. Kumar, et al. Annals of Medicine and Surgery 49 (2020) 37–40

Hospital readmission for cellulitis is also more common in older Table 1

people [18] particularly if there has been more than one prior episode Patient characteristics.
of cellulitis19. Age < 75 years Age 75 + years P-value
(N = 69) (N = 31)
2. Aims
Age mean(SD) 53.4 ( ± 14.2) 84.4 ( ± 5.8) < 0.001
Female n (%) 23 (33.3%) 16 (48.5%) 0.12
In this prospective study, we aimed to examine differences in risk BMI 36.0 ( ± 12.3) 28.3 ( ± 8.0) < 0.001
factors, clinical features, management, and outcomes of cellulitis be- Residential Aged Care 2 (2.9%) 8 (25.8%) 0.001
tween those 75 years or more and those less than 75 years admitted to a Facility
large metropolitan hospital. Mobility < 0.001
Mobile unaided 61 (88.4%) 13 41.9%)
Mobile with aid 8 (11.6%) 18 (58.1%)
3. Methods a
Risk factors OR (95% CI)
Dependent edema – 4.0 (1.3–12.6) 0.018
The study was conducted at Bankstown-Lidcombe Hospital, New Obesity (BMI > 30) – 0.3 (0.1–0.8) 0.012
Previous cellulitis – 0.3 (0.1–1.0) 0.044
South Wales, Australia from June 2014 to April 2015. The study was
Peripheral vascular – 3.1 (0.9–10.6) 0.079
approved by the South-Western Sydney Local Health District (SWSLHD) disease
Ethics Committee. Tinea pedis – 0.9 (0.2–4.1) 0.930
Between June 2014 and April 2015, potential patients were iden- Venous dermatitis – 0.5 (0.1–2.0) 0.302
tified through review of the Bankstown Hospital inpatient list three Comorbidities
Hypertension 16 (23.2%) 14 (45.2%) 0.035
times a week by a study investigator. We included all identified patients
AF 4 (5.8%) 11 (33.5%) < 0.001
aged 18 years or more with a diagnosis of cellulitis of the upper and/or Dementia 1 (1.4%) 7 (22.6%) 0.001
lower limb(s) and excluded patients with infected ulcers on presenta- Malignancy 1 (1.4%) 5 (16.1%) 0.010
tion, pregnant patients and those with post-operative wound infections. Diabetes 23 (31.9%) 6 (19.4%) 0.24
The patients were then stratified into an older group (aged 75 years IHD 11 (15.9%) 10 (32.3%) 0.11
CCF 10 (14.5%) 9 (29.0%) 0.10
or more) and a younger group (74 years or less) and were followed up DVT 3 (4.3%) 5 (16.1%) 0.10
during their admission and for a total of 28 days post completion of PE 2 (2.9%) 5 (16.1%) 0.072
intravenous antibiotics. We studied the over 75 years age group as that Steroid use last 3 3 (4.3%) 3 (9.7%) 0.37
this age group is more descriptive of the frail older cohort [19]. months
Data collected included basic demographics, clinical characteristics, a
Logistic regression – Chi-square = 17.868, p = 0.007, df = 6, Nagelkerke's
relevant investigations, treatment provided and clinical outcomes. The
R2 0.230; BMI body mass index; IHD ischaemic heart disease; AF atrial fri-
severity of cellulitis was rated using the Eron classification [20]. brillation; CCF congestive cardiac failure; DVT deep vein thrombosis; PE pul-
Data were analyzed with SPSS Version 24 and R version 3.3.1. Chi- monary embolism.
square test was used to compare proportions. Student's T-test was used
to compare differences in means for normally distributed variables. For similar in both groups.
non-normally distributed continuous variables, non-parametric test was Cellulitis presenting features such as pain, fever, chills and vital
used to assess differences in the ranked median scores. Logistic re- signs (temperature, heart rate and blood pressure) did not significantly
gression was used to assess statistically significant risk factors for cel- differ between the two groups. The severity of cellulitis, as defined by
lulitis in the older and younger age groups. Statistically significant re- the Eron classification [22] also did not differ between groups with the
sults were set at an alpha level of 0.05. The study is in line with the majority of patients having Eron Classes I and II (Table 2).
STROCSS guidelines [21]. The study also been registered on the re- Initial laboratory results revealed that older patients had lower
search registry UIN:researchregistry5125. hemoglobin [122.1 ( ± 16.4) vs 135.0 ( ± 19.4), p = 0.002] and al-
bumin [38.0 ( ± 47) vs 41.4 ( ± 4.1), p < 0.001] and higher urea
4. Results level [7.9 (5.8–12.4) vs 5.8 (4.8–8.4), p = 0.011] compared to their
younger counterparts. CRP white cell count (WCC) and positive rate of
One hundred and thirteen patients were identified during the study blood culture did not differ between the two groups (Table 2).
period and 100 patients (88.5%) consented to participate. Thirty-one Older inpatients presenting with cellulitis were less likely to be re-
(31.0%) patients were aged 75 years and older and 69 (69.0%) patients ferred to hospital in the home (HITH) antibiotic programs for comple-
were 74 years or less. tion of the course of intravenous antibiotics 32.3% vs 59.4% respec-
The mean age was 84.4 ± 5.8 years in the older group and tively, p = 0.012) compared to younger patients. The antibiotic choices
53.4 ± 14.2 years in the younger group. The older patients had lower did not differ between the two populations, these included Cephazolin,
BMI than their younger counterparts [28.3 ( ± 8.0) vs 36.0 ( ± 12.3) Flucloxacillin or Tazobactam-Piperacillin.
respectively, p < 0.001]. A higher proportion resided in residential Older patients with cellulitis were more likely to experience falls or
aged care facilities (25.8% vs 2.9% respectively, p = 0.001); and mo- decreased mobility (38.7% vs 15.9% respectively, p = 0.020) com-
bilized with walking aid(s) (58.1% vs 11.6% respectively, p < 0.001). pared to the younger group. (Table 3). Despite this, they had similar
(Table 1). LOS to their younger counterparts [10 (7–15) vs 8 (6–13) respectively,
A significantly higher proportion of older patients had documented p = 0.403]. There was one death in each group and the rates of ICU
hypertension (45.2% vs 23.2% respectively p = 0.035), atrial fibrilla- admission, surgical intervention and 28-day readmission were similar
tion (33.5% vs 5.8% respectively, p < 0.001), dementia (22.6% vs in the two groups.
1.4% respectively, p = 0.001) and malignancy (16.1% vs 1.4% re-
spectively, p = 0.010). (Table 1).
In terms of cellulitis risk factors, after controlling for potential 5. Discussion
confounders, older patients were more likely to have dependent edema
(OR 4.0 95%CI 1.3–12.6, p = 0.018); but less likely to be obese (OR In this study, we found that older people, despite being frailer than
0.3, 95%CI 0.1–1.0, p = 0.012) or had a prior history of cellulitis (OR their younger counterparts, had similar treatment outcomes after pre-
0.3, 95%CI 0.1–1.0, p = 0.044) than younger patients. The risk of senting to hospital with mild to moderate limb cellulitis.
peripheral vascular disease, tinea pedis and cutaneous dermatitis were In our study, most of the potential risk factors for cellulitis were

38
M. Kumar, et al. Annals of Medicine and Surgery 49 (2020) 37–40

Table 2 prevalence of these conditions in the older population rather than an

Clinical characteristics. association with cellulitis.
Age < 75 years Age 75 + years P-value Over 25% of older patients with cellulitis lived in residential aged
(N = 69) (N = 31) care facilities. This finding raised the opportunity for the provision of
ambulatory care antibiotic programs in aged care homes potentially
Duration of symptoms 3.0 (2.0–6.0) 4.0 (2.0–14.0) 0.23
avoiding the need for hospitalization for residents with cellulitis.
median (IQR)
Fever and chills 19 (27.5%) 6 (19.4%) 0.46
There were no significant differences in the clinical presentation of
Heart rate 92 ( ± 17) 89 ( ± 15) 0.384 cellulitis between the two age groups (i.e., duration of cellulitis symp-
BP – Systolic 136 ( ± 18) 143 ( ± 24) 0.158 toms, heart rate, blood pressure, temperature, white cell count, CRP
BP _Diastolic 75 ( ± 12) 72 ( ± 11) 0.227 and Eron severity classification). Atypical and blunted physiological
Temperature 37.4 ( ± 1.0) 37.3 ( ± 1.0) 0.796
response to infection with age has been documented in the literature
Pain score 0.055
Mild 0-3 32 (51.6%) 14 (53.8%) [23]. In severe sepsis, a reduced physiological response can lead to
Moderate 4-7 20 (32.3%) 12 (46.2%) rapid progression of sepsis [2,3,22].Our results did not support a
Severe 8-10 10 (16.1%) 0 blunted response to infection in older patients with cellulitis. We,
Pathology
however, did not have any patients with severe sepsis to examine the
Haemoglobin 135.0 ( ± 19.4) 122.1 ( ± 16.4) 0.002
White cell count 11.3 ( ± 5.1) 11.1 ( ± 6.0) 0.61
inflammatory response in more detail.
Albumin 41.4 ( ± 4.1) 38.0 ( ± 4.7) 0.001 In our study, older patients experienced more falls and impaired
Creatinine – median (IQR) 88 (76–106) 93 (71–131) 0.692 mobility during the admission for cellulitis compared to younger pa-
Urea – median (IQR) 5.8 (4.8–8.4) 7.9 (5.8–12.4) 0.011 tients. While these factors might have made their hospital discharge
CRP – median (IQR) 33 (15–117) 60 (9–133) 0.919
planning more complex, they did not translate into an increased hos-
Blood culture 29 (42.0%) 18 (58.1%) 0.071
Eron Classification 0.415 pital length of stay. Previously described risk factors affecting LOS in
Class I 10 (14.5%) 2 (6.5%) cellulitis (comprising of age, hypoalbuminemia, bacteremia, obesity, dia-
Class II 55 (79.7%) 26 (83.9%) betes mellitus, tachycardia, hypotension, leukocytosis, and elevated serum
Class III 4 (5.8%) 3 (9.7%) creatinine) [7,12–15], tended to be skewed towards age and hy-
Class IV 0 0
poalbuminemia for the older group and obesity for the younger group
BP blood pressure; CRP C reactive protein. in our study.
There were no statistically significant differences between the two
Table 3 groups in terms of mortality, ICU admission, and surgical intervention
Treatment, complications and outcomes. for cellulitis complications. The majority of patients in both groups had
Eron Class I or II cellulitis and did not sustain physiological decom-
Characteristics Age < 75 years Age 75 + years P-value
(N = 69) (N = 31)
pensations; however, in more severe cases of cellulitis, one would ex-
pect ageing physiology to sustain more physiological decompensations
Completed treatment via 41 (59.4%) 10 (32.3%) 0.012 which may then influence the above parameters.
HiTH A lower proportion of older inpatients discharged to HITH programs
Duration of IV antibiotic – 6 (4–8) 4 (2–9) 0.059
median (IQR)
might have been attributable to their medical comorbidities and func-
Length of hospital stay – 8 (6–13) 10 (7–15) 0.403 tional criteria not meeting HITH requirements. As such, additional
median (IQR) health resources may allow HITH programs to manage these complex
Antibiotics 0.121 patients but this would require further study.
Cephazolin 47 (51.1%) 16 (32.7%)
Unlike previous published literature [18], we did not find a sig-
Flucloxacillin 20 (21.7%) 12 (24.5%)
Tazobactam-piperacillin 7 (7.6%) 4 (8.2%) nificant difference in the 28-day readmission rate between the two age
Complications cohorts in our study. As the readmission rate was less than 5%, a study
DVT 0 0 1 with greater number of patients would have more power to detect small
PE 0 1 (3%) 0.31 differences in readmission rates.
Fall or decreased mobility 11 (15.9%) 12 (38.7%) 0.020
Nosocomial infection 1 (1.4%) 3 (9.7%) 0.087
One of the limitations of this study is the small sample size due to a
Delirium 1 (1.4%) 2 (6.5%) 0.23 short recruitment period; further study with a larger sample size would
Outcomes assist in validation of our findings. We decided to focus on inpatient
Death 1 (1.4%) 1 (3.0%) 0.531 cellulitis treatment; however a cellulitis management journey from
Needing surgical 3 (4.3%) 2 (6.5%) 0.644
hospital to community settings would have provided with a more
intervention
ICU admission 0 0 1 complete picture.
Readmission within 28 4 (5.8%) 0 0.308 As the number of older patients presenting with cellulitis increases
days as the population ages, it is important to note that for mild to moderate
Duration of IV antibiotic – 6 (4–8) 4 (2–9) 0.059 cellulitis, older patients perform just as well as younger patients with
median (IQR)
standard cellulitis treatments on clinical and care indicators.
Length of hospital stay – 8 (6–13) 10 (7–15) 0.403
median (IQR) We wish to confirm that there are no known conflicts of interest
Antibiotics 0.121 associated with this publication and there has been no financial or
Cephazolin 47 (51.1%) 16 (32.7%) person support for this work with any other people or organizations.
Flucloxacillin 20 (21.7%) 12 (24.5%)

HiTH hospital in the home; DVT deep vein thrombosis; PE pulmonary embo- Ethical approval
lism.
Ethics approval was taken from South-Western Sydney Local Health
similar in the older and younger age groups; however, older patients District (SWSLHD) Ethics Committee.
were more likely to have dependent edema and impaired mobility, and
less likely to be obese. Other conditions noted to be more common in
the older group were congestive cardiac failure, atrial fibrillation, de- Sources of funding
mentia and malignancy. We believe this finding reflected the higher
No funding was obtained for the research.

39
M. Kumar, et al. Annals of Medicine and Surgery 49 (2020) 37–40

Author contribution [5] C.G. Gunderson, R.A. Martinello, A systematic review of bacteremias in cellulitis
and erysipelas, J. Infect. 64 (2) (2012) 148–155, https://doi.org/10.1016/j.jinf.
2011.11.004.
Bin Soo Ong – Supervisor, study concept. [6] S. Esposito, S. Noviello, S. Leone, Epidemiology and microbiology of skin and soft
Manoj Kumar – Writing the paper. tissue infections, Curr. Opin. Infect. Dis. 29 (2) (2016) 109–115.
Huong Van Nguyen – writing the paper. [7] A. Dupuy, H. Benchikhi, J.C. Roujeau, P. Bernard, L. Vaillant, O. Chosidow, et al.,
Risk factors for erysipelas of the leg (cellulitis): case-control study, BMJ 318 (7198)
Clarence Yeong – Data collection. (1999) 1591–1594.
Vincent Ngian – data analysis, study concept. [8] S. Björnsdóttir, M. Gottfredsson, A.S. Thórisdóttir, G.B. Gunnarsson,
Caitlin Keighley – study concept, data collection. H. Ríkardsdóttir, I. Hilmarsdóttir, Risk factors for acute cellulitis of the lower limb:
a prospective case–control study, Clin. Infect. Dis. 41 (10) (2005) 1416–1422.
[9] M. Karppelin, T. Siljander, J. Vuopio-Varkila, J. Kere, H. Huhtala, R. Vuento, et al.,
Research registration number Factors predisposing to acute and recurrent bacterial non-necrotizing cellulitis in
hospitalized patients: a prospective case–control study, Clin. Microbiol. Infect. 16
(6) (2010) 729–734.
1. Name of the registry: Research Registry.
[10] J. Halpern, R. Holder, N.J. Langford, Ethnicity and other risk factors for acute lower
2. Unique Identifying number or registration ID: re- limb cellulitis: a U.K.-based prospective case–control study, Br. J. Dermatol. 158 (6)
searchregistry5125. (2008) 1288–1292.
3. Hyperlink to the registration (must be publicly accessible): [11] T.D. Kish, M.H. Chang, H.B. Fung, Treatment of skin and soft tissue infections in the
elderly: a review, Am. J. Geriatr. Pharmacother. 8 (6) (2010) 485–513.
https://www.researchregistry.com/browse-the-registry#home/ [12] A. Garg, J. Lavian, G. Lin, C. Sison, M. Oppenheim, B. Koo, Clinical characteristics
registrationdetails/5d7a4d440ba5620010271dc1/ associated with days to discharge among patients admitted with a primary diag-
nosis of lower limb cellulitis, J. Am. Acad. Dermatol. 76 (4) (2017) 626–631.
[13] S.C. Morpeth, S.T. Chambers, K. Gallagher, C. Frampton, A.D. Pithie, Lower limb
Guarantor cellulitis: features associated with length of hospital stay, J. Infect. 52 (1) (2006)
23–29.
Dr Manoj Kumar. [14] M. Figtree, P. Konecny, Z. Jennings, C. Goh, S.A. Krilis, S. Miyakis, Risk stratifi-
cation and outcome of cellulitis admitted to hospital, J. Infect. 60 (6) (2010)
431–439.
Consent [15] M. Theofiles, J. Maxson, L. Herges, A. Marcelin, K.B. Angstman, Cellulitis in obesity:
adverse outcomes affected by increases in body mass index, J. Prim. Care
Community Health 6 (4) (2015) 233–238.
All authors have consented to publication. [16] R. Tan, D.J. Newberry, G.J. Arts, M.E. Onwuamaegbu, The design, characteristics
and predictors of mortality in the North of England cellulitis treatment assessment
Provenance and peer review (NECTA), Int. J. Clin. Pract. 61 (11) (2007) 1889–1893.
[17] J. Carratalà, B. Rosón, N. Fernández-Sabé, E. Shaw, O. del Rio, A. Rivera, et al.,
Factors associated with complications and mortality in adult patients hospitalized
Not commissioned, externally peer reviewed. for infectious cellulitis, Eur. J. Clin. Microbiol. Infect. Dis. 22 (3) (2003) 151–157.
[18] A. Garg, J. Lavian, G. Lin, C. Sison, M. Oppenheim, B. Koo, Clinical factors asso-
Declaration of competing interest ciated with readmission among patients with lower limb cellulitis, Dermatology 233
(1) (2017) 58–63.
[19] Y. Ouchi, H. Rakugi, H. Arai, M. Akishita, H. Ito, K. Toba, I. Kai, Joint committee of
No conflict of interest from any of the authors. Japan gerontological society (JGLS) and Japan geriatrics society (JGS) on the de-
finition and classification of the elderly, Redefining the elderly as aged 75 years and
older: Proposal from the joint Committee of Japan gerontological Society and the Japan
References geriatrics society, Geriatr. Gerontol. Int. 17 (7) (2017 Jul) 1045–1047.
[20] M. Karppelin, T. Siljander, J. Aittoniemi, M. Hurme, R. Huttunen, H. Huhtala, et al.,
[1] National Health Performance Authority Analysis of Admitted Patients Care National Predictors of recurrent cellulitis in five years. Clinical risk factors and the role of
Minimal Data Set 2013-2014 and Australian Bureau of Statistics Estimated Resident PTX3 and CRP, J. Infect. 70 (5) (2015) 467–473.
Population 30 June 2013. [21] R.A. Agha, M.R. Borrelli, M. Vella-Baldacchino, R. Thavayogan, D.P. Orgill, for the
[2] D.J. Anderson, K.S. Kaye, Skin and soft tissue infections in older adults, Clin. STROCCS Group, The STROCCS statement: strengthening the reporting of cohort
Geriatr. Med. 23 (3) (2007) 595–613. studies in Surgery, Int. J. Surg. 46 (2017) 198–202.
[3] S. Laube, Skin infections and ageing (2004, Ageing Res. Rev. 3 (1) (2004) 69–89. [22] L.J. Eron, Infections of skin and soft tissues: outcome of a classification scheme,
[4] Q.Y. Weng, A.B. Raff, J.M. Cohen, N. Gunasekera, J.P. Okhovat, P. Vedak, et al., Clin. Infect. Dis. 31 (A432) (2000) 287.
Costs and consequences associated with misdiagnosed lower extremity cellulitis, [23] D.C. Norman, D. Grahn, T.T. Yoshikawa, Fever and aging, J. Am. Geriatr. Soc. 33
JAMA Dermatol. 153 (2) (2017) 141–146. (1985) 859.

40