The Final Bow? Increased Mortality in Elderly Patients With Acute Bacterial Parotitis: A Retrospective Observational Study

Acute bacterial parotitis (ABP) is a severe and potentially life-threatening infection of the parotid gland. It is characterised by pain, swelling and erythema in the parotid region with purulent discharge from the parotid duct [1, 2]. While ABP can occur at any age, it primarily affects elderly and immunocompromised individuals [2-4]. ABP is thought to be precipitated by reduced salivary flow secondary to dehydration, with subsequent retrograde migration of infection (most commonly Staphylococcus aureus) along the parotid duct [2]. Other common risk factors for xerostomia include medications like diuretics and anticholinergics and diseases like Sjogren's syndrome [1, 2]. Diagnosis is largely clinical, with imaging, like ultrasound or computed tomography (CT) scans being reserved for cases with suspected ductal obstruction or abscess formation [2]. The mainstay of ABP management involves broad-spectrum antibiotics. For mild cases, oral therapy may be appropriate, but for most hospitalised patients, intravenous antibiotics are utilised [1, 2]. In cases where abscesses form, needle aspiration or surgical drainage may be indicated [5]. Protocols regarding the parent team responsible for ABP patients vary worldwide. In the UK, most patients are admitted under Ear Nose and Throat (ENT), general medicine or care of the elderly (COTE) teams based on the referrer's perception of the most appropriate team. This approach lacks standardised processes to guide selection, resulting in variability in the quality of care delivered. In our experience in a district general hospital, we observed that patients with ABP were frequently referred to ENT with concerns regarding potential parotid abscess, however, this was rarely found. In such situations, the predominantly elderly patients would often require a prolonged hospital admission on a surgical ward without requiring surgical intervention and without access to the rehabilitation options more readily available in the COTE wards. In addition, unfortunately two patients passed away under our care after the episode of ABP had resolved. This prompted us to establish the incidence of parotid abscess and the overall mortality associated with ABP for patients admitted to our centre. We hope this data helps establish the optimum management of these patients and improve outcomes. Data analysis was conducted using retrospective analysis of a prospectively collected database of patients admitted under ENT at a single UK district general hospital between September 2019 and December 2023. Clinical information was obtained retrospectively from clinical notes. The eligibility criteria were patients aged 18 years and above who were admitted with acute parotitis. Eighteen episodes of ABP in 16 patients met our eligibility criteria with two of these episodes being recurrences. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline was the reporting guideline of choice for manuscript preparation. This article is based on a local clinical practice audit aimed at improving patient care. As such, informed consent and research ethics committee approval were not required. The average age of the patients admitted with ABP was 72 years (range 27–97). 11 patients (69%) were above 65 years, while 5 (31%) were under 65 years. There were 12 women and 4 men. All patients above 65 years were polymorbid. The average Charlson Comorbidity Index Score was 5.3, indicating a 2%–21% estimated 10 years survival rate [6]. ENT was the parent team for 81% of the patients, with the remainder being managed by general medical teams. Three patients had parotid pus swabs taken. Two grew Staphylococcus aureus, while the third grew Klebsiella pneumoniae. Five patients had blood cultures taken. Four specimens were normal while one grew Staphylococcus epidermidis (likely contaminant) and Escherichia coli. Escherichia coli bacteraemia was felt to be clinically significant and originated from ABP. The average CRP was 175 (range 11–515) while average WCC was 16.7 (range 6.7–43.9). Patients baseline observations on admission were not available retrospectively to determine the presence of sepsis on admission. Seven (39%) patients had imaging conducted. The most common was ultrasound (US) with 3 (18%) cases using US only and 2 (12.5%) patients having both CT and US scans. Two (12.5%) patients underwent CT scans only. None of the patients had a parotid abscess or required surgical intervention. One (6.3%) 27-year-old patient required ultrasound guided aspiration for a lymph node abscess which was performed by radiology during diagnostic ultrasound. Median length of inpatient stay was 5 days (range 2–84). We were unable to ascertain if a new or increased package of care was required for patients from this retrospective study or whether this delayed hospital discharge. While the average number of months to death was 5.1 months, death during the ABP admission varied significantly between age groups. The overall mortality rate during admission was 13%. No patients under 65 years died during their admission. However, in the patients over 65 years, the mortality rate during admission was 18%. Overall mortality rates were also calculated for all patients at 1 month, 1 and 1.5 years from discharge. The overall mortality was 25% at 1 month, 50% at 12 months and escalated to 80% at 18 months from discharge. Notably, all deaths occurred exclusively in the above 65-year group, leading to mortality rates of 36% at 1 month, 75% at 12 months and 100% at 18 months from discharge, as shown in Table 1 and Figure 1. Death was due to many different reasons. Only one patient had parotitis as the main cause of death (Table 2). Overall, our review reveals that no patients within our sample had a parotid abscess or required surgical intervention. Although one patient required radiological intervention, they were young and the indication was for a cervical lymph node abscess, rather than a parotid abscess. While parotid abscesses are thought to be rare, one study found that the incidence of confirmed parotid abscesses in patients with acute infectious parotid disease was 19.7% [7]. This is in stark contrast to our experience. Due to our limited study size, we are unable to conclude the true incidence of parotid abscesses. Larger studies are required to establish the complication rate of this. When parotid abscesses do occur, some suggest that antibiotics and rehydration remain as the cornerstone of treatment and that incision and drainage should be reserved as a salvage option due to its high complication rate [2, 5]. This suggests that ABP could be appropriately managed by medical teams in the early stages, with referral to ENT as needed in those with treatment failure. This would also allow for more holistic care of these polymorbid patients. In our study, ABP was typically not the main cause of mortality. In addition, mortality exclusively occurred in patients above 65 years. There is currently limited literature on the relationship between ABP and mortality in the elderly. There are a few anecdotal reports from geriatric, ENT and palliative care teams, with the largest study having 22 patients and a 27% mortality rate 2 weeks after diagnosis of ABP [3]. Other smaller studies report mortality rates of 78% within 1 month of symptom onset and 80% within 5 weeks of symptom onset [4, 8] (Table 3). All studies recognise parotitis as a marker of poor prognosis. Similarly, neck of femur fractures are linked with high mortality, with studies indicating a 1-year mortality rate of 14%–58% with standard care. Factors related to mortality included increasing age and a Charlson Comorbidity Score of 4 and higher [9, 10]. In our study, the average patient age was 72 years, and the average Charlson Comorbidity Score was 5.3. A high comorbidity burden often leads to polypharmacy, which increases the risk of xerostomia—a key predisposing factor for ABP. This suggests that ABP should be viewed as a surrogate marker for significant frailty and elevated mortality in the 12–18 months post-admission, similar to the case with neck of femur fractures. In contrast to the studies shown in Table 3, where patients were primarily admitted for a separate illness, most patients within our study were admitted primarily due to ABP, as might be expected from an ENT-led study. Our study is also unique in its inclusion of patients beyond the geriatric population, highlighting the differing impact of the disease in a younger cohort. Limitations of this study include its small size, due to the rarity of this condition. Additionally, the study only includes patients referred to ENT. As referrals are often made due to concerns regarding complications of ABP, this study is therefore subject to selection bias which may explain the high associated mortality. It is plausible that milder cases are managed in the community and have different mortality rates. Given the retrospective nature of this study, there is also the possibility of incomplete handover data, resulting in some patients being inadvertently excluded from the study. In conclusion, ABP rarely requires surgical intervention and is considered a poor prognostic indicator, especially in the context of patients over 65 years. Clinicians should consider these factors when determining the appropriate specialty for managing ABP, ensuring optimal patient care. We hope our results will inform future clinical decisions and prompt further research, such as an age-matched cohort study. S.L. and N.W. designed the work. S.L., N.W. and R.V. acquired and analysed data. S.L., N.W. and R.T. drafted, revised and approved the manuscript. S.L., N.W., R.V. and R.T. agree to be accountable for all aspects of the work. This article is based on a local clinical practice audit aimed at improving patient care. As such, informed consent and research ethics committee approval were not required. Due to our significant findings, we hope our results help inform future clinician decisions, ensuring that patients receive optimal care. The authors declare no conflicts of interest. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/coa.14289. The data that support the findings of this study are available on request from the corresponding author, N.W. The data are not publicly available due to patient confidentiality.