Should we reconsider cefazolin for treating staphylococcal meningitis? A retrospective analysis of cefazolin and cloxacillin cerebrospinal fluid levels in patients treated for staphylococcal meningitis

OBJECTIVES

To assess the meningeal penetration of cefazolin and cloxacillin in individuals treated for methicillin-susceptible staphylococcal meningitis.

METHODS

We retrospectively identified individuals treated for Staphylococcus meningitis with measurements of cefazolin or cloxacillin concentrations in cerebrospinal fluid (CSF) using a validated assay of liquid chromatography coupled with mass spectrometry at the Nantes University Hospital between January 2009 and October 2019. Staphylococcus meningitis was defined by a compatible clinical presentation and a microbiological confirmation (positive CSF culture or positive specific PCR). Medical charts were retrospectively reviewed to collect microbiological and clinical data, and to assess therapeutic success.

RESULTS

Among the 17 included individuals, eight (47%) were treated with cefazolin and nine (53%) with cloxacillin. Median daily dosages of cefazolin and cloxacillin were 8 g (range 6-12 g) and 12 g (range 10-13 g), respectively. Cefazolin and cloxacillin were mainly administered by continuous infusion. Eleven individuals (65%) were men, median (interquartile range (IQR)) age was 54 years (50; 70), 14 (82%) had postoperative meningitis and 3 (18%) had haematogenous meningitis. Median (IQR) antibiotic CSF concentrations were 2.8 mg/L (2.1; 5.2) and 0.66 mg/L (0.5; 0.9) for cefazolin and cloxacillin groups, respectively. Cloxacillin was discontinued in two individuals for therapeutic failure.

CONCLUSIONS

Patients with staphylococcal meningitis treated with high-dose continuous intravenous infusion of cefazolin achieved therapeutic concentrations in CSF. Cefazolin appears to be a therapeutic candidate that should be properly evaluated in this indication.

Quelle surveillance des infections associées aux soins en réanimation en 2018 ?

La surveillance des infections associées aux soins (IAS) est prioritaire en réanimation, secteur à haut risque du fait de l’état critique des patients et de leur exposition aux dispositifs invasifs. Elle présente un triple objectif : décrire l’épidémiologie et l’incidence des IAS ; évaluer l’impact des actions de prévention ou de contrôle et alerter face à une épidémie ou des phénomènes émergents. Cette surveillance des IAS peut être réalisée selon une méthodologie interne, définie par l’établissement, ou intégrée à un réseau de surveillance. L’intérêt de la surveillance pour la prévention des IAS en réanimation n’est plus à démontrer, mais la surveillance manuelle reste chronophage pour les cliniciens et les équipes d’hygiène, limitant ainsi le temps dédié à la prévention de ces infections. La surveillance automatisée apparaît aujourd’hui comme un outil intéressant, tant par ses performances que par le gain de temps qu’elle représente pour les équipes. Plusieurs éléments sont primordiaux pour obtenir des résultats fiables : la nécessité d’une harmonisation des définitions et des méthodes de surveillance ; la mise à disposition d’outils informatiques performants pour faciliter le suivi des patients ; le leadership des réanimateurs dans la surveillance. Cet article fait le point sur les méthodes de surveillance des IAS utilisées aujourd’hui en réanimation, l’intérêt de la mise en place de cette surveillance épidémiologique ainsi que la fiabilité des données recueillies et, enfin, les avantages du développement d’une surveillance semi-automatisée ou automatisée des IAS dans ce secteur.

Influence of surgical bleeding on the relationship between admission coagulopathy and risk of massive transfusion: lesson from 704 severe trauma patients

BACKGROUND

This study hypothesized that the relationship between early coagulopathy and massive transfusion (MT) in trauma was highly dependent on the presence of surgical bleeding.

METHODS

Consecutive severe trauma patients admitted to our institution over a 4-year period were included in this retrospective study. Surgical bleeding was defined as an injury requiring an invasive endovascular or surgical haemostatic procedure. The ability of prothrombin time ratio (PTr) and activated partial thromboplastin time ratio (aPTTr) to predict MT (≥10 units of packed red blood cells during the first 24 h) was determined by ROC curves. The strength of association and interaction between PTr, surgical bleeding and MT was assessed using a logistic regression analysis.

RESULTS

Among the 704 patients included (ISS 21·0 ± 16·2), MT rate was higher in patients with surgical bleeding than in those with no surgical bleeding (47% vs. 5%; P < 0·001). The global performance of PTr and aPTTr to predict MT was only fair in our study population (AUCs 0·83 and 0·81). MT rate was widely higher in the surgical bleeding group whatever the severity of coagulopathy (P < 0·001). PTr was found to be significantly associated with TM [PTr ≥ 1·5, OR 23·6 (95% CI 13·4-41·7); PTr 1·2-1·5, OR 3·0 (95% CI 1·7-5·3)]. Corresponding ORs were reduced after adjusting for the surgical bleeding: 12·1 (95% CI 6·5-22·5) and 2·1 (95% CI 1·2-4·0), respectively. However, no significant interaction was found regression models.

CONCLUSION

The strength of association between MT and coagulation status on admission was found strongly influenced by surgical bleeding. The admission coagulopathy monitoring in trauma patients without considering the surgical bleeding does not allow a reliable determination of MT probability.

Blood pressure monitoring during arrhythmia: agreement between automated brachial cuff and intra-arterial measurements

BACKGROUND

Since arrhythmia induces irregular pulse waves, it is widely considered to cause flawed oscillometric brachial cuff measurements of blood pressure (BP). However, strong data are lacking. We assessed whether the agreement of oscillometric measurements with intra-arterial measurements is worse during arrhythmia than during regular rhythm.

METHODS

Among patients of three intensive care units (ICUs), a prospective comparison of three pairs of intra-arterial and oscillometric BP readings was performed among patients with arrhythmia and an arterial line already present. After each inclusion in the arrhythmia group, one patient with regular rhythm was included as a control. International Organization for Standardization (ISO) standard validation required a mean bias <5 (sd 8) mm Hg.

RESULTS

In 135 patients with arrhythmia, the agreement between oscillometric and intra-arterial measurements of systolic, diastolic and mean BP was similar to that observed in 136 patients with regular rhythm: for mean BP, similar mean bias [-0.1 (sd 5.2) and 1.9 (sd 5.9) mm Hg]. In both groups, the ISO standard was satisfied for mean and diastolic BP, but not for systolic BP (sd >10 mm Hg) in our ICU population. The ability of oscillometry to detect hypotension (systolic BP <90 mm Hg or mean BP <65 mm Hg), response to therapy (>10% increase in mean BP after cardiovascular intervention) and hypertension (systolic BP >140 mm Hg) was good and similar during arrhythmia and regular rhythm (respective areas under the receiver operating characteristic curves ranging from 0.89 to 0.96, arrhythmia vs regular rhythm between-group comparisons all associated with P>0.3).

CONCLUSIONS

Contrary to widespread belief, arrhythmia did not cause flawed automated brachial cuff measurements.

A haemoperitoneum does not indicate active bleeding in the peritoneum in 50% of hypotensive blunt trauma patients: a study of 110 severe trauma patients

BACKGROUND

We hypothesised that in blunt trauma patients with haemodynamic instability and haemoperitoneum on hospital admission, the haemorrhagic source may not be confined to the peritoneum. The purpose of this study was to describe the incidence and location of bleeding source in this population.

METHODS

The charts of trauma patients admitted consecutively between January 2005 and January 2010 to our level I Regional Trauma Centre were reviewed retrospectively. All hypotensive patients presenting a haemoperitoneum on admission were included. Hypotension was defined by a systolic blood pressure ≤ 90 mmHg. The haemoperitoneum was quantified on CT images or from operative reports as moderate (Federle score<3 or between 200 and 500 ml) or large (Federle score ≥ 3 or >500 ml). Active bleeding (AB) was defined as injury requiring a surgical or radiologic haemostatic procedure, regardless of origin (peritoneal (PAB) or extraperitoneal (EPAB)).

RESULTS

Of 1079 patients admitted for severe trauma, 110 patients met the inclusion criteria. Seventy-eight (71%) were male, mean age 35.3 (SD 19) years and mean ISS 36.5 (SD 20.5). Among the 91 patients who had AB, 37 patients (41%) had PAB, 34 (37%) had EPAB and 20 had both (22%). Forty-eight (53%) of them had moderate haemoperitoneum and 43 (47%) had large haemoperitoneum. A large haemoperitoneum had positive predictive value for PAB of 88% (95% CI 75-95%) and negative predictive value of 65% (95% CI 49-79%). The corresponding values in the subgroup of patients with EPAB were 65% (95% CI 38-86%) and 76% (95% CI 59-88%).

CONCLUSION

Haemoperitoneum was associated with PAB in only 52% of hypotensive blunt trauma patients and 63% of bleeding patients. In contrast, 59% of bleeding patients had at least one EPAB. The screening of a haemoperitoneum as a marker of active haemorrhagic source may be confusing and lead to misdiagnosis and inappropriate strategy. Clinician should exclude carefully the presence of any EPAB explaining haemorrhagic shock, before to decide haemostatic treatment.

Brachial cuff measurements of blood pressure during passive leg raising for fluid responsiveness prediction

OBJECTIVE

The passive leg raising maneuver (PLR) for fluid responsiveness testing relies on cardiac output (CO) measurements or invasive measurements of arterial pressure (AP) whereas the initial hemodynamic management during shock is often based solely on brachial cuff measurements. We assessed PLR-induced changes in noninvasive oscillometric readings to predict fluid responsiveness.

STUDY DESIGN

Multicentre interventional study.

PATIENTS AND METHODS

In ICU sedated patients with circulatory failure, AP (invasive and noninvasive readings) and CO measurements were performed before, during PLR (trunk supine, not modified) and after 500-mL volume expansion. Areas under the ROC curves (AUC) were determined for fluid responsiveness (>10% volume expansion-induced increase in CO) prediction.

RESULTS

In 112 patients (19% with arrhythmia), changes in noninvasive systolic AP during PLR (noninvasiveΔ(PLR)SAP) only predicted fluid responsiveness (cutoff 17%, n=21, positive likelihood ratio [LR] of 26 [18-38]), not unresponsiveness. If PLR-induced change in central venous pressure (CVP) was at least of 2 mm Hg (n=60), suggesting that PLR succeeded in altering cardiac preload, noninvasiveΔ(PLR)SAP performance was good: AUC of 0.94 [0.85-0.98], positive and negative LRs of 5.7 [4.6-6.8] and 0.07 [0.009-0.5], respectively, for a cutoff of 9%. Of note, invasive AP-derived indices did not outperform noninvasiveΔ(PLR)SAP.

CONCLUSION

Regardless of CVP (i.e., during "blind PLR"), noninvasiveΔ(PLR)SAP more than 17% reliably identified fluid responders. During "CVP-guided PLR", in case of sufficient change in CVP, noninvasiveΔ(PLR)SAP performed better (cutoff of 9%). These findings, in sedated patients who had already undergone volume expansion and/or catecholamines, have to be verified during the early phase of circulatory failure (before an arterial line and/or a CO measuring device is placed).

Povidone iodine: features of critical systemic absorption

Povidone iodine (PI), a skin antiseptic, is sometimes used internally but this procedure exposes to potentially lethal iodine absorption. Indeed, a 41-year-old woman, with no relevant medical history, developed a transient hypotension, anuric renal failure, hemolysis, coagulopathy and uterine infarction after intra-uterine injection of PI as a dye to check the fallopian tube patency (hydrotubation). Iodemia peaked at 6929 μg/dL (normal range 3.4-8.0 μg/dL), and decreased over the 9 days of renal replacement therapy. Extreme caution should be exercised when PI is in contact with a mucosa, the early recognition of iodine toxicity being of utmost importance to rapidly prompt renal replacement therapy. The main purpose of this report is to highlight the clinical features of PI absorption, whatever the route of administration.

Concurrent Graves disease thyrotoxicosis and myasthenia gravis: the treatment of the former may dangerously reveal the latter

Autoimmune thyroxicosis and myasthenia gravis are often associated. In both diseases, clinical features may include neuromuscular weakness, making their distinction challenging. We report a patient with known Graves disease who presented with generalised fatigue, initially attributed solely to thyrotoxicosis, and who experienced severe respiratory failure linked to associated myasthenia gravis that was unmasked by medication used in the perioperative management of his thyroxicosis. Anaesthetists should always consider myasthenia gravis in cases of hyperthyroidism presenting with neuromuscular features.