Impact of MALDI-TOF implementation and antimicrobial stewardship programs on the optimization of antibiotic therapy in Staphylococcus aureus bloodstream infections

Abstract

Introduction

Rapid pathogen identification is key in the management of bloodstream infections. Antimicrobial stewardship programs (ASPs) have been shown to reduce resistance, adverse effects, and healthcare costs. MALDI-TOF MS enables fast and accurate microbial identification. This study evaluates the impact of its implementation on antibiotic therapy adequacy and mortality in Staphylococcus aureus bacteremia.

Methods

A quasi-experimental pre-post study was conducted at a university hospital, covering two periods: June 2015/April 2017 (pre) and June 2021/April 2023 (post). The impact of MALDI-TOF on antibiotic adequacy and mortality was analyzed.

Results

A total of 176 episodes were included (85 pre, 91 post). No significant epidemiological differences were observed between groups, although higher rates of sepsis and septic shock were noted in the post phase. MALDI-TOF was associated with improved antibiotic adjustment (3.51 vs. 1.75 days; P < .001). A non-significant trend toward lower early mortality was observed in the post-intervention group (P = .06). Mortality was associated with higher comorbidity (Charlson index), sepsis/septic shock, and MRSA bacteremia.

Conclusion

The implementation of MALDI-TOF, in conjunction with a multidisciplinary ASP, enhances early adjustment of empirical antibiotic therapy and may be associated with reduced early mortality in S. aureus bacteremia. Incorporating rapid diagnostics, especially for MRSA detection, is a key strategy to improve clinical outcomes.

Resumen

Introducción

La identificación rápida del patógeno es clave en el tratamiento de las bacteriemias. Los programas de optimización del uso de antibióticos (PROA) han demostrado reducir resistencias, efectos adversos y costes. MALDI-TOF MS permite una identificación microbiana rápida y precisa. Este estudio evalúa el impacto de su implementación en la adecuación del tratamiento antibiótico y la mortalidad en bacteriemias por Staphylococcus aureus.

Métodos

Estudio cuasiexperimental pre-post en un hospital universitario, con 2 periodos: junio 2015/abril 2017 (pre) y junio 2021/abril 2023 (post). Se analizó el impacto de MALDI-TOF en la adecuación antibiótica y la mortalidad.

Resultados

Se incluyeron 176 episodios (85 pre, 91 post). No hubo diferencias epidemiológicas significativas entre grupos, aunque se observó mayor frecuencia de sepsis y shock séptico en la fase post. MALDI-TOF se asoció a un mejor ajuste antibiótico (3,51 vs. 1,75 días; p < 0,001). Se observó una tendencia no significativa a menor mortalidad precoz en el grupo post (p = 0,06). La mortalidad se asoció con mayor comorbilidad (índice de Charlson), sepsis/shock séptico y bacteriemia por SARM.

Conclusión

La implementación de MALDI-TOF, junto a un PROA multidisciplinar, optimiza el ajuste precoz de la antibioterapia empírica, y podría estar asociada a una reducción de la mortalidad en las fases iniciales de la bacteriemia por Staphylococcus aureus. Implementar diagnósticos rápidos, especialmente para detectar SARM, es una estrategia clave para mejorar los resultados clínicos.

Introduction

Despite advances and improvements in antimicrobial therapy, among all infections, bacteraemia remains one of the leading causes of morbidity and mortality.¹ Early antibiotic therapy can improve patient prognosis and reduce mortality rates and healthcare costs.^2,3 However, even when sensitivity results are available, it is not uncommon for delays in the provision of appropriate treatment due to the lack of specific specialised interventions.³

Antimicrobial stewardship programmes (ASP) play a crucial role in improving the appropriateness of antibiotic therapy while minimising the unintended consequences of antimicrobial use, such as toxicity, selection of pathogenic organisms such as Clostridioides difficile and the emergence of resistance.⁴

ASP have been shown to be effective in reducing antimicrobial resistance and the associated economic burden, especially in settings with limited resources.⁵ In addition, compared to mere reporting of results, they allow for early review of microbiological results and real-time intervention on patient management.⁶ Rapid identification of the pathogen causing bacteraemia is essential, as a delay in microbiological diagnosis can lead to suboptimal empirical treatment which does not cover resistant strains such as methicillin-resistant Staphylococcus aureus (MRSA)⁷ and unnecessary exposure to broad-spectrum antibiotics, contributing to the development of resistance and an increased risk of adverse events.⁸

Microbiological diagnostic stewardship (MIDS) programmes, designed to improve the quality of diagnosis by ensuring that methods are the most accurate, reliable and rapid, enable more effective clinical intervention.⁹ Rapid microbial identification from positive blood cultures by mass spectrometry, known as matrix-assisted laser desorption-ionisation time-of-flight mass spectrometry (MALDI-TOF MS), makes it possible for clinicians to adapt antibiotic therapy early before the antimicrobial susceptibility result is available.¹⁰ However, there is not yet systematic application of this technique in the routine practice of all microbiology laboratories.

Recent studies have shown that the implementation of rapid diagnostic tests, such as MALDI-TOF, combined with the intervention of ASP teams, is associated with significant clinical improvements such as a reduction in mortality rates and reduced time to optimal treatment.^11,12 However, the specific clinical impact of this intervention on S. aureus bacteraemia in settings with an ASP already in place remains unclear.

In this scenario, our hypothesis is that rapid microbial identification by MALDI-TOF from positive blood cultures, with direct communication to the clinician, would enable early optimisation of empirical antibiotic therapy before the antimicrobial susceptibility result is available, thus improving cure rates and reducing mortality rates in patients with S. aureus bacteraemia.

Methods

The study was conducted in a 582-bed tertiary public hospital covering a health area with a population of about 312,000. The hospital has a Microbiology service that works 24 h a day, seven days a week, and has had MALDI-TOF MS technology in its blood culture workflow since May 2017. We designed a retrospective, single-centre, pre-post quasi-experimental study to collect clinical and analytical variables from patients with S. aureus bacteraemia registered in the microbiology department of our centre during the period June 2015 to April 2017, this period considered prior to the implementation of the microbial identification system through MALDI-TOF, and during the period June 2021 to April 2023, these two full years considered as being representative of when the MALDI-TOF technology had been introduced and consolidated within the workflow in the microbiology laboratory. In the previous period (2015–2017), bacterial identification was performed by conventional methods based on biochemical tests on subculture and direct antibiogram of positive blood culture. During both periods, the Microbiology service was continuously active (24/7), including the reading and interpretation of blood cultures. The communication of results during working hours (Monday to Friday, 8 a.m. to 3 p.m.) is done through the ASP team (composed of specialists in microbiology, infectious diseases and hospital pharmacy), which reviews the prescriptions, acting directly or advising the treating specialist in person, by telephone and/or electronically through recommendations in the electronic medical record. Outside working hours, the microbiology laboratory directly notifies the on-call treating specialist or, in their absence, the internal medicine/infectious diseases team by telephone.

We included patients with S. aureus bacteraemia during the periods June 2015 to April 2017, and June 2021 to April 2023 who met all inclusion criteria (first episode of monomicrobial bacteraemia due to S. aureus) and none of the exclusion criteria (paediatric patients, defined as under 15 years of age, absence of clinical data in their electronic medical records, such as date and antibiotic prescribed, patient comorbidities, place of acquisition or outcome at 30 days, and episodes of polymicrobial bacteraemia).

A sample size of 80 patients per group was estimated to detect significant differences in the correct adaptation of targeted antibiotic regimens between the two diagnostic techniques; using an odds ratio of 2.5, with a confidence level of 95% and a power of 80%, assuming a loss of 5%, non-probabilistic consecutive sampling was applied.

The presence of immunosuppression was defined as the presence of innate or acquired immunodeficiency, immunosuppressive treatments (corticosteroid therapy >3 months, chemotherapy, biologic drugs), history of organ or bone marrow transplantation. Risk for MRSA infection was considered to be the patient having come from residential care, prolonged hospitalisation in patients >65 years of age with multiple disorders, previous antibiotic therapy in the last three months, patient on haemodialysis, stay in intensive care units or invasive procedures, mucositis in neutropenic patients or skin ulcers or burn patients, as well as previous colonisation or infection by MRSA. Severity was categorised into isolated bacteraemia, sepsis (defined as infection with 2 or 3 points on the qSOFA scale),¹³ or septic shock (presence of sepsis together with the need for vasopressors for mean arterial pressure 65 mmHg or lactate >2 mmol/l after resuscitation with fluids).

Empirical treatment was defined as antibiotics administered when there was clinical suspicion of infection, before microbiological results were available. Targeted treatment was considered to be the treatment initiated after receiving the first available microbiological information: Gram staining in the pre-MALDI-TOF period, and identification by MALDI-TOF in the post-implementation period.

The criteria for categorising the appropriateness of antibiotic therapy as correct, improvable or incorrect were applied following the action protocols created by our hospital's ASP group and available to all hospital professionals on the centre's intranet. It was defined as correct if it coincided with the recommendations of the protocol based on suspected infectious focus, site of acquisition and severity, incorrect if the therapy did not cover isolation, and improvable if the antibiotic therapy required some adjustment either to broaden the antibiotic spectrum or to rationalise it. The appropriateness of empirical and targeted treatment was determined on the basis of the treatment prescribed in the electronic medical record, taking into account the time at which it was recorded by the responsible physician. The hospital's electronic prescription system is supervised in real time by the Pharmacy service, so that any modification made between the hours of 8 a.m. and midnight automatically generates the preparation of the treatment and its immediate dispatch to the corresponding ward. At night (from midnight to 8 a.m.), once the prescription has been issued, the nursing staff contacts the Pharmacy to request the medication.

The primary outcome variable was clinical status after the episode of bacteraemia categorised by cure (resolution of symptoms, negative blood cultures at least 48 h after antibiotic therapy, haemodynamic stability and completion of antibiotic therapy with no recurrence of infection observed at a follow-up at least 30 days after completion), overall death (death within 30 days as a result of the episode of bacteraemia) and persistent bacteraemia.

Statistical analysis was carried out using the statistical software Stata® version 14.2.

For sample selection, pairwise matching was used with a 1:1 ratio between the two periods adjusted for age, gender, immunosuppression and S. aureus methicillin resistance. To compare differences between groups, Chi-square or Fisher's exact test for categorical samples and Student's t-test or Mann-Whitney U test were used to compare differences between means and medians. To assess associations between risk factors and outcomes, a logistic regression analysis was performed. This model allowed calculation of the adjusted odds ratio (OR), together with its corresponding 95% confidence interval (95% CI). Kaplan-Meier survival curves were used to estimate the probability of survival and differences between groups were compared using the log-rank test, with differences with a p-value of less than 0.05 being considered statistically significant.

From a total of 312 episodes of S. aureus bacteraemia, after pairwise matching with a 1:1 ratio between the two periods adjusted for age, gender, immunosuppression and S. aureus methicillin resistance, 195 blood cultures were included, 176 of which met all inclusion criteria and none of the exclusion criteria. A total of 85 bacteraemia episodes were collected in the period prior to the introduction of the MALDI-TOF MS system for microbial identification and 91 episodes were included in the subsequent stage (Fig. 1).

Figure 1.

Flowchart of the bacteraemia episodes included in the study.

Results

Comparison of epidemiological characteristics

There were no differences between groups in terms of gender, mean age, Charlson index, immunosuppression, source of patient, site of acquisition or source of bacteraemia (Table 1). There was a trend towards a higher number of episodes of sepsis and septic shock in the post-MALDI-TOF MS group, but no statistically significant differences. There was no difference between the two groups with respect to having a central venous catheter (CVC) in situ. There was a trend towards a higher proportion of patients with a urological device in the pre-MALDI-TOF MS group with 32 cases (37.65%), compared to the post-MALDI-TOF MS group, which had 23 cases (25.27%). The main source of the bacteraemia was similar in both groups, with a predominance of catheter, respiratory and skin and soft tissue.

Table 1.

Demographic and clinical data of the patients included in the study.

Variable	Pre MALDI-TOF MS phase period (%)	Post MALDI-TOF MS period (%)	p-value
	n = 85	n = 91
Mean age (SD)	66.03 (14.84)	66.47 (15.67)	0.738
Gender			0.844
Male	55 (64.71)	55 (60.42)
Female	30 (35.29)	36 (39.56)
Charlson Comorbidity Index; mean (SD)	3.47 (2.62)	3.6 (2.78)	0.85
Charlson Comorbidity Index >3	54 (63.53)	49 (53.85)	0.328
Immunosuppression	17 (20)	18 (19.78)	0.971
Department			0.557
Accident and Emergency	41 (48.24)	48 (52.75)
Internal medicine	8 (9.41)	5 (5.49)
Intensive Care Unit	7 (8.24)	3 (3.30)
Nephrology	6 (7.06)	5 (5.49)
Haematology	5 (5.88)	2 (2.20)
Oncology	4 (4.71)	7 (7.69)
Cardiology	3 (3.53)	2 (2.20)
General surgery	2 (2.35)	4 (4.40)
Trauma and Orthopaedics	2 (2.35)	1 (1.10)
Respiratory Medicine	2 (2.35)	2 (2.20)
Neurology	2 (2.35)	8 (8.79)
Endocrinology	1 (1.18)	1 (1.10)
Gastroenterology	1 (1.18)	3 (3.30)
Neurosurgery	1 (1.18)
Acquisition			0.640
Community-acquired	57 (67.06)	64 (70.33)
Nosocomial	28 (32.94)	27 (29.67)
Severity			0.850
Isolated bacteraemia	53 (62.35)	53 (58.24)
Sepsis	19 (22.35)	24 (24.18)
Septic shock	13 (15.29)	16 (17.58)
Source of bacteraemia			0.375
Catheter	19 (22.35)	18 (19.78)
Respiratory	16 (18.82)	15 (16.48)
Skin and soft tissue	17 (20)	14 (15.38)
Unknown	13 (15.29)	19 (20.88)
Endocarditis	5 (5.88)	8 (8.79)
Abdominal	5 (5.88)	1 (1.10)
Urinary	4 (4.71)	6 (6.59)
CNS	4 (4.71)	2 (2.20)
Joint-related	2 (2.35)	6 (6.59)
Hepatobiliary	0 (0)	2 (2.20)
MRSA bacteraemia	19 (22.35)	18 (19.78)	0.703
Previous MRSA colonisation	2 (2.35)	4 (4.40)	0.313
MRSA risk factors	28 (32.94)	47 (51.65)	0.012
Prior antibiotic therapy	11 (12.94)	16 (17.58)	0.394
History of skin and soft-tissue infection	3 (3.53)	7 (7.69)	0.233
Central venous catheter	31 (36.47)	27 (29.67)	0.318
Urological device	32 (37.65)	23 (25.27)	0.068

CNS: central nervous system; MALDI-TOF MS: matrix-assisted laser desorption-ionisation time-of-flight mass spectrometry; MRSA: methicillin-resistant Staphylococcus aureus; SD: standard deviation.

Significant differences were found with respect to risk factors for MRSA in the post-MALDI-TOF group, although the rates of MRSA colonisations and bacteraemia were similar between the two groups. The rest of the variables are shown in Table 1.

Adequacy and use of antibiotic therapy

The proportions of patients who received correct, improvable or incorrect empirical and targeted antibiotic therapy according to the detection method in each period are shown in Table 2. The proportion of correct treatments was similar in both study periods, although there was a trend towards a higher proportion of correct targeted treatments after the implementation of the MALDI-TOF MS system (p: 0.065).

Table 2.

Distribution by group according to appropriateness of empirical and targeted antibiotic therapy in both study stages.

	Pre MALDI-TOF MS	Post MALDI-TOF MS
Empirical antibiotic treatment
Correct, n (%)	57 (67.05)	53 (54.95)
Improvable, n (%)	14 (16.47)	27 (29.67)
Incorrect, n (%)	10 (11.76)	13 (14.29)
Targeted antibiotic therapy
Correct, n (%)	75 (88.24)	91 (100)
Improvable, n (%)	3 (3.52)	0 (0)
Incorrect, n (%)	0 (0)	0 (0)

MALDI-TOF MS: matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry.

Appendix B Figs. A.1 and A.2 in the Supplementary material show the distribution of empirical and targeted antibiotic therapy respectively used in both phases of the study. In the post-MALDI-TOF MS period, piperacillin-tazobactam, linezolid and daptomycin predominate compared to the pre-implementation period. Regarding the use of targeted antibiotic therapy according to the detection method, vancomycin and amoxicillin-clavulanic acid were most commonly used in the pre-MALDI-TOF MS period, with daptomycin and cloxacillin being the predominant antibiotic therapy in the post-MALDI-TOF MS period.

The mean time to correction of empirical antibiotic therapy was 3.51± 2.15 days in the period prior to implementation of the MALDI-TOF MS identification system versus 1.75euro± 1.65 days in the post-MALDI-TOF MS period, with statistically significant differences (p < 0.001).

Hospital admission and outcome

Fig. 2 shows the differences in clinical outcome between the two study periods. No significant differences were identified in cure (p: 0.997), persistent bacteraemia (p: 0.999) or 30-day mortality after the episode of bacteraemia (p: 0.676).

Figure 2.

Outcome of the bacteraemia in each of the periods of the study.

Most of the patients cured at the end of the episode were patients without sepsis or septic shock criteria in both groups (pre-MALDI-TOF MS: 47 [48.18%] and post-MALDI-TOF MS: 44 [48.35%]; p: 0.775). Most of the patients with septic shock were those with the highest mortality rates in both study periods (pre-MALDI-TOF MS: 8 [9.41%] and post-MALDI-TOF MS 10 [10.9%]; p: 0.683). The mean number of days of hospitalisation was 21.49euro± 22.11 during the pre-MALDI-TOF MS system period versus 21.19euro± 20.47 days in the post-implementation period, with a p: 0.911.

Mortality and associated risk factors

Table 3 shows the clinical and demographic characteristics compared between patient groups according to 30-day mortality for each of our study periods (pre- and post-MALDI-TOF MS), recording the magnitude of the effect on mortality of the significant variables associated with mortality in both intervention groups. The most notable are the effect on mortality of methicillin-resistant S. aureus bacteraemia (OR: 3.32 and 2.68 in pre- and post-MALDI-TOF MS groups respectively), and the presence of urological shunt devices (OR: 5 and 2.72 in pre- and post-MALDI-TOF groups, respectively). In a logistic regression model adjusted for severity, bladder catheterisation was not significantly associated with mortality (OR 1.45, 95% CI 0.62–3.39, p = 0.395). In contrast, both sepsis (OR: 3.67; 95% CI: 1.50–8.98; p = 0.004) and septic shock (OR: 9.31; 95% CI: 3.21–27.05; p < 0.001) were associated with a significantly increased risk of dying.

Table 3.

Mortality rates by clinical-epidemiological characteristics and study period.

	Pre MALDI-TOF MS			Post MALDI-TOF MS
	Did not die	Died	p-value	Did not die	Died	p-value
Mean age (SD)	65.71 (13.31)	69.76 (16.47)	0.259	64 (15.56)	74.16 (13.79)	0.005
Gender			0.173			0.140
Male	44 (51.76)	11 (12.94)		45 (49.45)	10 (10.99)
Female	20 (23.53)	10 (11.76)		20 (22)	16 (16.58)
Charlson Comorbidity Index; mean (SD)	3.39 (2.32)	4.85 (2.85)	0.02	3.21 (2.61)	5.2 (2.55)	0.002
Charlson Comorbidity Index >3	38 (44.71)	16 (18.82)	0.165	30 (32.97)	19 (20.88)	0.215
Immunosuppression	13 (15.29)	4 (4.71)	0.900	13 (14.29)	5 (5.49)	0.899
Acquisition			0.265			0.255
Community-acquired	45 (52.94)	12 (14.12)		47 (51.64)	17 (18.68)
Nosocomial	19 (22.35)	9 (10.59)		19 (20.88)	8 (8.79)
Severity			<0.001			<0.001
Isolated bacteraemia	48 (56.47)	5 (5.88)		48 (52.75)	5 (5.49)
Sepsis	8 (9.41)	11 (12.94)		10 (10.99)	14 (15.38)
Septic shock	5 (5.88)	8 (9.41)		6 (6.59)	10 (10.99)
MRSA bacteraemia	8 (9.41)	11 (12.94)	0.029	7 (7.69)	11 (12.09)	0.038
Previous MRSA colonisation	1 (1.18)	1 (1.18)	0.401	3 (3.3)	1 (1.09)	0.502
MRSA risk factors	19 (22.35)	9 (10.59)	0.265	27 (29.67)	20 (21.98)	0.365
Prior antibiotic therapy	7 (8.23)	4 (4.70)	0.350	8 (8.79)	8 (8.79)	0.550
History of skin and soft-tissue infection	2 (2.35)	1 (1.18)	0.724	4 (4.39)	3 (3.29)	0.522
Central venous catheter	26 (30.59)	5 (5.88)	0.165	16 (17.58)	11 (12.09)	0.265
Urological device	14 (16.47)	18 (21.18)	0.002	9 (9.89)	14 (15.38)	0.002

CNS: central nervous system; MALDI-TOF MS: matrix-assisted laser desorption-ionisation time-of-flight mass spectrometry; MRSA: methicillin-resistant Staphylococcus aureus.

Fig. 3 shows the Kaplan-Meier survival curves for the pre-MALDI-TOF MS (blue) and post-MALDI-TOF MS (red) periods. A higher mortality rate is observed in the pre-MALDI-TOF MS group in the first days, with a more rapid decline in survival compared to the post-MALDI-TOF MS group. However, towards the end of follow-up, the curves tend to converge, suggesting that differences in mortality rates decrease over time. Analysis using the log-rank test showed a difference in survival close to statistical significance (p = 0.06). Although the conventional threshold for statistical significance is not reached, the value obtained suggests that the implementation of MALDI-TOF could be associated with an improvement in survival, particularly in the first days, although this difference is not conclusive.

Figure 3.

Kaplan-Meier survival curve showing the effect of time to adequate provision of antibiotic therapy on mortality rates in both study periods.

Discussion

ASP have demonstrated significant improvements in hospitals where they have been implemented, achieving both clinical benefits and reductions in healthcare costs.⁵ There are different strategies for achieving these improvements, such as prioritisation of opportunities for improvement based on principles of antibiotic use, on care services, on drugs or on infectious syndromes.¹⁴ This area includes the MIDS programmes, which promote action aimed at optimising the use of diagnostic techniques, contributing to appropriate therapeutic and clinical decision-making.⁹ The use of MALDI-TOF MS technology for direct identification of positive blood cultures has been shown to substantially reduce the time, compared to conventional methods where the result can be delayed by 24−48 h, enabling early adjustment of empirical therapy.¹⁵

In our study, the high level of comorbidity represented by the validated Charlson scale, where more than 60% of patients had a score higher than 3, was similar to other studies.^16,17 The most common clinical focus was catheter-related bacteraemia, followed by respiratory and skin and soft tissue foci. Correct empirical treatment predominated over improvable or incorrect regimens in both phases. However, there was a notable a trend towards more appropriate regimens thanks to the information received through MALDI-TOF MS identification. These data are similar to those already published by Osthoff et al., although with some differences in the study design, with antimicrobial therapy evaluated within 24 h of positivity, and the results of Torres et al., although they had a more global approach, not only including S. aureus and focusing on oncology/haematology units and intensive care units.^15,18 In addition, an ASP team specialised in early antibiotic therapy review and adjustment was already in place at our centre and this may reduce external validity in other healthcare settings. One example of this is the work published by Yong Duk et al. where, with a very similar design, although in an environment without an ASP team, no positive outcomes were found in response to the early information obtained by the MALDI-TOF MS system.¹⁹ In any event, despite early detection using the MALDI-TOF MS system, the mortality rate increased in patients with greater comorbidities assessed with validated scales such as the Charlson index, this index being an independent predictor of mortality.²⁰ The presence of a urological bypass device as a risk factor associated with mortality was not confirmed in the logistic regression analysis adjusting for severity, so this observation was probably due to the fact that patients with urological devices were those who showed greater severity, requiring continuous monitoring of diuresis. The presence of sepsis or septic shock is strongly associated with worse outcomes with mortality rates of 38–86 %.^21–23 Although definitions of sepsis and septic shock vary in the literature, the data collected in other published studies are similar to those we obtained in our study.

With regard to the mortality rate associated with MRSA compared to methicillin-sensitive (MSSA), we found data in the literature that reinforce these findings.²⁴ There could potentially be a number of possible causes for this association. Firstly, specifically related to pathogen characteristics, virulence factors associated with the staphylococcal cassette chromosome mec(SCCmec), with SCCmec type II being an independent predictor of mortality in the cohort study published by Chen SY et al.²³ Secondly, differences in empirical vancomycin prescribing may explain the diversity of results, as vancomycin has a slower in vitro bactericidal activity, especially in infections with high inoculum size and variable tissue distribution.²⁵ Lastly, it is possible that in MRSA infections, methicillin resistance per se is a confounding factor and actually reflects the impact of host-related factors such as comorbidities. In our study, we did not consider the effect of rapid detection of PBP2a by immunochromatography, currently standard practice for the early detection of methicillin resistance,²⁶ as it was not performed systematically during the period studied and we specifically sought to analyse the impact of MALDI-TOF and the ASP intervention.

In terms of the importance of time to adequate antibiotic therapy as a survival factor, multiple publications have found that delaying adequate treatment has a detrimental effect.^2,27 A recent study by Fidalgo et al. in a cohort of more than 6000 episodes of bacteraemia showed that delayed reporting of clinically relevant microbiological findings is associated with a significant increase in mortality rates in patients with bacteraemia.²⁸ However, other studies have failed to detect differences in mortality rates, or only in selected patient groups with a high APACHE II, with complicated S. aureus bacteraemia²⁹ or in patients with a non-eradicable focus compared to eradicable foci.³⁰ Different definitions of appropriate empirical treatment, the groups of patients studied or possible biases in the studies could explain these discrepancies. In any event, it is reasonable to assume that early prescribing of appropriate antibiotics may be beneficial in selected patient groups.

This study has several limitations. Firstly, this is a quasi-experimental study without randomisation, which was why matching techniques were applied between the two groups according to key variables such as age, gender, immunosuppression and S. aureus methicillin resistance. Secondly, the experience is limited to a single centre, which reduces the external validity of the study and limits the ability to generalise the conclusions to other hospital settings with different structures. Thirdly, the implementation of rapid identification by MALDI-TOF MS took place in a context where a consolidated ASP team was already in place, with early intervention in the management of bacteraemia, which might have reduced the observable margin for improvement and, consequently, made it difficult to identify statistically significant differences. Lastly, as detailed in the methods section, the time of treatment initiation was established based on the electronic prescription record, which could introduce slight imprecision with regard to the exact time of antibiotic administration.

In summary, our findings reinforce the clinical utility of rapid diagnosis by MALDI-TOF MS in combination with established ASP programmes to optimise antibiotic therapy in S. aureus bacteraemia. Although the improvement in treatment adequacy did not translate into a statistically significant reduction in the overall mortality rate, the observed benefits in early adequacy and the trend in early survival underline the importance of these interventions. Future multicentre studies, with larger sample sizes and prospective design, will be key to confirming these results and more precisely defining their impact on clinical outcomes.

Ethical considerations

This study was reviewed and approved by the Independent Ethics Committee of Hospital Clínico Universitario de Valencia (order 2024/197). The study was conducted in accordance with the principles of the Declaration of Helsinki and local regulations for research in humans. CBS has a Rio Hortega contract (CM23/00282), funded by the Instituto de Salud Carlos III (ISCIII) [Carlos III Health Institute].

Funding

This work has not been financed by any additional budget.