**Evaluation of Pediatric Surgical Site Infections Associated with Colorectal Surgeries at an Academic Children's Hospital**

#### **Kimberly Pough 1,2,\*, Rima Bhakta <sup>3</sup> , Holly Maples 1,3, Michele Honeycutt <sup>4</sup> and Vini Vijayan <sup>5</sup>**


Received: 12 March 2020; Accepted: 5 April 2020; Published: 9 April 2020

**Abstract:** Appropriate use of antibiotic prophylaxis (AP) is a key measure for the prevention of surgical site infections (SSI) in colorectal surgeries; however, despite the presence of national and international guidelines, compliance with AP recommendations remains low. The purpose of this study is to evaluate compliance with recommendations for the use of AP in children undergoing colorectal surgeries and to evaluate the effectiveness of antibiotics in the prevention of SSI. We collected demographic and clinical characteristics of patients who underwent colorectal surgeries, as well as microbiological and antimicrobial susceptibility data for patients who developed SSI. AP data were collected and compared with national guidelines. Antibiotic dosing and duration were most frequently in concordance with national guidelines, while antibiotic timing and selection had the lowest rates of compliance. Twelve of the 192 colorectal procedures evaluated resulted in SSI. Only 2 of the 12 children with SSI received appropriate AP for all four categories evaluated. Eight cases that resulted in SSI were due to organisms not covered by the recommended AP. We identified multiple areas for the improvement of AP in children undergoing colorectal surgery. A multidisciplinary approach to development of standardized protocols, educational interventions, and EHR-based algorithms may facilitate or improve appropriate AP use.

**Keywords:** colorectal surgery; pediatric; surgical prophylaxis; antibiotic prophylaxis; surgical site infections

#### **1. Introduction**

According to the Centers for Disease Control and Prevention (CDC), surgical site infections (SSI) are infections that occur at or near the surgical incision within 30 days of a procedure, or 90 days for specified procedures [1]. These infections occur in approximately 2–5% of patients undergoing inpatient surgery in the United States, and account for approximately 20% of healthcare-associated infections in adults as well as children [2]. SSI are associated with high morbidity and mortality rates, and increased durations of hospitalization and healthcare costs [2–4].

Colorectal surgeries are associated with a higher rate of SSI than for other kinds of surgeries, ranging from 5% to 45% due to exposure to the increased bacterial load in the colon and the rectum [3–7]. Current guidelines published by the CDC and the Healthcare Infection Control Practices Advisory Committee for the Prevention of Surgical Site Infection recommend appropriate utilization of systemic antibiotic prophylaxis (AP) within a surgical bundle as a key measure to prevent SSI among patients undergoing colorectal surgeries [8]. Appropriate AP in colorectal procedures is based mainly on 4 principles: (1) correct antibiotic selection; (2) correct dose; (3) timing of administration, including appropriate re-dosing for extended procedures; and (4) discontinuation of antibiotics when the procedure is completed and surgical site is closed, or no more than 24 h post-operatively. The effectiveness of AP in the prevention of SSI is well established. In 2016, the World Health Organization published evidence-based recommendations regarding the use of AP in the prevention of SSI [1,2,9–12]. However, despite the presence of international and national guidelines, compliance with AP for surgical procedures has been staggeringly low among patients undergoing colorectal procedures [13,14].

Clinical evidence in support of AP for the reduction of infectious complications following colorectal surgery is derived almost exclusively from adult literature. There are no well-controlled studies evaluating the efficacy of AP and compliance with surgical AP in children undergoing colorectal procedures. However, as children and adults have similar fecal bacterial concentration and microbiological profiles, there is little reason to suspect that current guidelines would not be adequate for children [3,15,16].

The purpose of this study is to evaluate the compliance of surgeons to national recommendations for use of AP in children undergoing colorectal surgeries with particular regard to antibiotic selection, dose, timing prior to incision and intraoperative re-dosing, and duration of postoperative antibiotic use and is to evaluate the effectiveness of antibiotics in the prevention of SSI in children undergoing colorectal surgical procedures.

#### **2. Materials and Methods**

#### *2.1. Study Design and Setting*

We performed a retrospective cohort study at Arkansas Children's Hospital (ACH) in Little Rock, Arkansas. ACH is a 336-bed academic teaching hospital and serves as the largest children's hospital in Arkansas. This project was conducted in accordance with the Declaration of Helsinki, and the institutional review board of the University of Arkansas for Medical Sciences approved this study on November 21, 2017 (Protocol number 207,026), using expedited review procedures.

#### *2.2. Study Population*

The study population included all pediatric patients <18 years of age who underwent a colorectal procedure at ACH from 1 January 2015 to 31 December 2016. We excluded surgeries that were performed as the direct result of trauma and those surgeries without anesthesia records available. A list of potentially eligible patients was provided by the hospital infection prevention team. Patients were identified through chart review of colorectal procedures and application of the standardized National Healthcare Safety Network (NHSN) definitions for SSI at the time of reporting. SSI were defined and reported to NHSN for each procedure, including all SSI types: superficial, deep, and organ space. The wound class system used in NHSN is adapted from the American College of Surgeons wound classification schema and includes Clean, Clean-Contaminated, Contaminated, and Dirty/Infected [17]. SSI were determined through prospective surveillance by four infection control practitioners who are certified in infection prevention with 4–26 years of experience. This known subset of children provided the opportunity for assessment of antibiotic prophylaxis utilization.

#### *2.3. Data Collection*/*Study Procedures*

We performed a comprehensive review of medical records by using a standardized data collection instrument to identify demographic information and clinical characteristics of patients who underwent colorectal surgeries at ACH. Perioperative antibiotic use, dose, timing of first administration, and duration of prophylaxis were collected and compared with the American Society of Health-System Pharmacists (ASHP) guidelines for appropriate use of antibiotics for surgical prophylaxis [9]. Microbiological and antimicrobial susceptibility data for patients who developed an SSI post-operatively were obtained from our institution's microbiology laboratory for the 2-year time period.

#### *2.4. Definitions*

Based on the ASHP guidelines, appropriate antibiotic selection for colorectal procedures was defined as those using one of the following regimens: (1) cefazolin and metronidazole, (2) ceftriaxone and metronidazole, (3) cefoxitin, (4) cefotetan, (5) ampicillin-sulbactam, or (6) ertapenem [9]. Alternative regimens for patients with a beta-lactam allergy included clindamycin or metronidazole with an aminoglycoside, aztreonam, or a fluoroquinolone [9]. Vancomycin could be used in the place of clindamycin for patients with a beta-lactam allergy [9]. Inappropriate antibiotic selection was defined as any other regimen administered preoperatively for the purposes of AP. Antibiotic dose was considered appropriate if the administered dose was within 10% of the guideline recommended dose.

Antibiotic timing was categorized as appropriate or inappropriate. Appropriate timing was defined as administration of the first dose of antibiotics within 60 min prior to surgical incision. However, given the pharmacokinetics of fluoroquinolones and vancomycin, timing of 120 min prior to incision was deemed appropriate for those antibiotics. Antibiotics not administered during these time periods were considered as inappropriate timing.

Re-dosing interval was assessed from the time of administration of the preoperative dose of the antibiotic and deemed appropriate when given within two half-lives of the agent administered, and deemed inappropriate when not administered or if there was a delay in administration. Continuation of AP for >24 h after surgery without an infectious indication was deemed as inappropriate duration.

#### *2.5. Statistical Analysis*

We performed descriptive analyses of the above variables by using SPSS version 24. Testing of proportions was performed by using a χ2 or Fisher exact test as appropriate. All reported *p* values are 2-tailed and were considered significant if *p* < 0.05.

#### **3. Results**

We evaluated 208 colorectal surgical procedures, of which 192 children met the inclusion criteria. Sixteen patients were excluded due to lack of anesthesia records. Of the 192 surgeries performed, 12 (6%) met the NHSN criteria for a surgical site infection; the overall SSI rate was 6.25 per 100 surgical procedures.

The median age of all patients was 4.9 months (range, 0–17.7 years), and 113 (59%) were male. Fifteen (8%) children were overweight or obese. One hundred seventy-five (91%) surgeries were categorized as scheduled or elective, and 17 (9%) were urgent or emergent. The median duration of surgery was 92 min (range, 20–579 min). The median duration of hospitalization was 13 days (range, 1–511 days). Of the 192 patients, 62 (32%) were hospitalized at least once in the previous year.

The types of surgeries most frequently performed included colorectal resection (44%), ostomy formation/revision (35%), ostomy closure (34%), exploratory laparotomy (28%), and small bowel resection (17%); most patients required multiple surgery types during their procedures. Table 1 shows the demographic, clinical, and surgical characteristics of the patients.


**Table 1.** Demographic and clinical characteristics of study cohort.

#### *3.1. Assessment of AP Compliance*

The results of compliance with antimicrobial prophylaxis are shown in Figure 1. Appropriate antibiotic dosing and duration had the highest incidence of compliance at 65% and 64% of cases, respectively. Antibiotic timing and selection had the highest rates of non-compliance at 56% and 64% of encounters being non-compliant, respectively.

#### **Antibiotic Appropriateness**

**Figure 1.** Appropriateness of antibiotic prophylaxis for children undergoing colorectal surgery as compared to national guideline recommendations.

Antibiotic selection was found to be in concordance with both local and national recommendations in 36% of the cases (69/192). Combination of cefazolin and metronidazole was the most common appropriately used antibiotic regimen, accounting for 26% of all surgical cases. The most common inappropriate antibiotic regimens selected included cefazolin monotherapy and a combination of vancomycin with piperacillin-tazobactam. Vancomycin alone was administered to three patients, and metronidazole alone was administered to one patient. Anaerobic coverage was not included in the antibiotic regimen in 62% of patients. Thirty-five percent of patients for whom AP was not selected appropriately were on scheduled antibiotics for an infection prior to surgery, and hence, AP was perceived to be not indicated per surgical documentation. One patient did not receive any AP, and 6% of children that did not receive appropriate AP had a documented beta-lactam allergy.

– With regard to antibiotic timing, 56% (107/192) of patients received AP outside of the recommended administration time. Of the 107 inappropriately timed antibiotics, 24 (22%) were due to vancomycin administration beyond the optimal time window prior to incision (range, 97–1144 min); 9 (8%) were due to emergency procedures. We found that 24 (22%) cases of inappropriate antibiotic timing were due to delay in the administration of metronidazole following the administration of cefazolin, ceftriaxone, or a fluoroquinolone. Of the 50 patients who were already on scheduled antibiotics prior to surgery, one received antibiotics at the appropriate time prior to incision, and three were appropriately re-dosed intraoperatively.

We found that dosing of AP was inappropriate in 68/192 (35%) of our patients. Dosing errors were noted most frequently for metronidazole; 71 patients in our cohort received metronidazole preoperatively, of which 29 (41%) received a higher dose than recommended, while 13 (18%) patients received a suboptimal dose of metronidazole. Of the 23 patients that required re-dosing of antibiotics, only 8 (35%) were re-dosed appropriately. The median surgical duration for procedures that required re-dosing was 177 min (range, 52–577 min).

– AP duration was inappropriate in 70/192 (36%) cases. The duration of antibiotics after surgical procedure in patients whose post-operative prophylaxis was inappropriately prolonged was a median of 48.63 h (range, 31.33–182.62 h).

– Overall, noncompliance with all four elements of antimicrobial prophylaxis was 44% among the 192 cases (Table 2).


**Table 2.** Appropriateness of antibiotic prophylaxis in children undergoing colorectal surgery.

Note: For dual combinations, both antibiotics had to be appropriate.

#### *3.2. Surgical Site Infections*

Twelve children (6%) in our cohort developed SSI following colorectal surgery. Of these, 5 were superficial incisional, 2 were deep incisional, and 5 were organ/space infections. The percentages of clean-contaminated, contaminated, and dirty wounds in patients who developed infection were 25%, 17%, and 58%, respectively. Of the surgical cases resulting in SSI, 42% were emergent cases. Seventeen percent of infections occurred in patients who were obese and 25% occurred in patients who were premature. The median duration of surgery in cases resulting in SSI was 112.5 min (range, 76–206 min). Cases involving bowel resections accounted for 83% of all SSI. –

Of the 12 patients with SSI, only two children received the correct AP for all four categories evaluated including selection, time, dose, and duration. Antibiotics were inappropriately selected in 4/12 (33%) children who developed an SSI. AP timing, duration, and dosing were inappropriate in 6/12 (50%), 5/12 (42%), and 2/12 (17%) cases, respectively.

The organisms isolated in patients with SSI were methicillin-susceptible *Staphylococcus aureus* (MSSA), methicillin-resistant *Staphylococcus aureus* (MRSA), *Escherichia coli, Enterococcus faecalis, Pseudomonas aeruginosa, Klebsiella pneumoniae, Enterobacter cloacae, Candida albicans, Candida tropicalis*, and *Candida glabrata* (Figure 2).

**Figure 2.** Organisms isolated in 12 children with a surgical site infection following colorectal surgery.

Of the 10 children with an SSI wherein an organism was identified, 8 (80%) were not covered by the recommended AP. Of these 8 cases, 3 (38%) were due to *Candida* sp., and 5 (63%) were due to organisms that were resistant to the standard AP.

#### **4. Discussion**

We found lack of compliance with national guidelines in all four facets of AP in children undergoing colorectal procedures at our institution. Appropriate antibiotic selection and timing had the highest incidence of non-compliance, but we also identified non-compliance with antibiotic dosing and duration.

Antibiotic selection had the highest rate of non-compliance in our study with the correct antibiotic being chosen in only 36% of children undergoing colorectal surgeries. At our institution, the choice of AP is at the discretion of the surgeon or anesthesiologist. Lack of familiarity with the national guidelines for AP may be a barrier to appropriate antibiotic selection. In a study published by Friedman et al., excessively broad-spectrum antibiotics were chosen for clean operations [18]. This finding was similar to that seen in our study wherein concern for serious or severe infection prompted surgeons to unnecessarily choose broader spectrum antibiotics, thereby placing patients at risk for antibiotic resistance and fungal infections. The use of clinical decision support pathways and order sets that are incorporated into the electronic medical record may help guide antibiotic selection and prevent antibiotic overuse [14,19]. These order sets should be developed with the input of pharmacists and include optimal dosing for the chosen antibiotic, thereby potentially overcoming the AP dosing issues noted in our study. A multi-disciplinary approach including pharmacy, surgeons, nursing, and anesthesia for the development of the clinical decision support pathway may help to shed light on different perspectives of patient care and hold all members of the patient care team accountable for ensuring appropriate use of AP [14,19,20].

Nearly 60% of inappropriate AP in our study was due to incorrect timing. The ASHP guidelines recommend to administer antibiotics within one hour prior to surgical incision, or within 120 min for specific antibiotics [9]. We noted that when dual antibiotics were selected for AP, the second antibiotic, most often metronidazole, was either delayed or administered at or after the time of incision. The reason for this is unclear, but lack of familiarity with the pharmacokinetics of antibiotics may be a contributing factor. Tan et al. reported that AP was perceived as a low priority when compared to the administration of anesthetics among surgeons and anesthesiologists, and that this likely influenced the timing of AP [21]. Incorporation of AP into the routine operating room workflow and administration of prophylaxis in the pre-operative area rather than in the operating room may ensure complete infusion of antibiotics prior to incision. As anesthesiologists play a critical role in postoperative infection control, the delegation of AP administration to the anesthesiology team should be considered. Nemeth et al. evaluated use of a verbal AP reminder in the surgical time-out process, but found that this intervention did not improve timeliness of administration of AP [21]. Nair et al. demonstrated the effectiveness of direct email feedback, antibiotic compliance reports, and real time alerts in improving antibiotic timing [22].

Our findings of variation in AP practices are similar to that of other studies evaluating the use of AP in pediatric surgical patients. Donà et al. noted variability in antibiotic prescribing for AP in their single-center study that evaluated the use of AP in children undergoing surgical procedures. The authors found that in the pre-intervention group, antibiotic selection was inappropriate in 51% of cases, and antibiotics were continued for a prolonged duration in 54.9% of cases [23]. Implementation of a clinical pathway proved to be a useful tool and led to a statistically significant improvement in the selection and duration of AP in pediatric patients; however, there still remained room for improvement of AP compliance in the post-intervention group [23]. Sandora et al. evaluated the national appropriateness of AP in children undergoing common surgical procedures using the Pediatric Health Information System database, and they noted significant variation in the use of AP across the 31 institutions submitting data [24]. AP was considered to be appropriate in only 64.6% of all cases in the study, with an inter-hospital variation ranging from 47.3% appropriateness to 84.4%. The authors noted that AP was commonly administered, even in cases for which AP was not indicated, revealing a significant overuse of antibiotics despite the presence of national guidelines and well described risks of antibiotic associated adverse reactions and secondary infections, such as *Clostridioides di*ffi*cile*

infection [24]. They also concluded that the lack of pediatric guidelines for AP may have impacted this finding of variability in AP practices between hospitals [24]. Additionally, while it is commonly inferred that the colonic composition in children is similar to that of adults, studies have demonstrated differences between the pediatric and adult gut microbiome [25,26]. Furthermore, the disproportionate differences in chronic conditions and comorbidities between children and adults may lend to different post-operative SSI risks when comparing these two populations [24]. Considering these differences, surgeons may be less inclined to extrapolate the adult guidelines to their pediatric patients.

AP was effective in the prevention of SSI in our study and only 6% developed an SSI. Of those children that developed an SSI, 80% were due to infections that were not covered by standard AP, 25% were premature infants, and 17% were in obese patients. It is well known that antibiotic overuse is frequent in neonates and significant variability exists in their use. Neonates are, therefore, at risk for antibiotic resistant organisms. Currently, there are limited data on appropriate surgical AP specific to neonates and AP in this population are based on adult guidelines [27]. Considering the unique microbiome of neonates and the morbidity associated with SSI in neonates, larger studies are warranted to determine effective AP in this particular population, as conventional AP may not be optimal. Two of the seven obese patients in this cohort developed an SSI. Patients who are obese commonly undergo longer operative times and are at risk for increased complications and prolonged hospitalizations following surgery [28]. Furthermore, the lack of data regarding antibiotic dose adjustments in obesity lends to the concern that these patients may not have adequate serum drug concentrations when standard doses of AP are utilized. Based on our small study, these special populations may benefit from a more tailored AP regimen.

This study has several limitations. This was a single-center study and, hence, the findings may not be generalizable to all pediatric surgical settings. Due to the retrospective study design, we were limited to information reported in the patients' medical records; therefore, findings may have been misclassified if the data points were not completely recorded in the chart. The application of a clinical chart review may not have captured all facets of SSI documentation. We did not evaluate the use of oral antibiotics for mechanical bowel prophylaxis prior to elective colorectal procedures, so it is unclear if those practices were impactful in preventing SSI in our cohort. Finally, SSI cases were identified using a list provided by our infection preventionists using NHSN criteria; however, cases may not have been captured if cultures were not obtained despite objective signs leading to clinical suspicion of infection, such as fever or wound drainage.

#### **5. Conclusions**

In this study, we have identified multiple areas for improvement regarding the administration of AP in children undergoing colorectal surgeries. Lack of compliance with national guidelines for AP in children undergoing colorectal surgeries was high. A multidisciplinary approach to the development of standardized protocols, educational interventions, and EHR-based algorithms may facilitate or improve appropriate AP use. Special populations, such as neonates and obese children, may benefit from a tailored regimen for AP, as these children may be at risk for SSI due to organisms not covered by conventional AP regimens. Our findings indicate the need for larger studies to investigate optimal AP choices in special populations and to determine interventions to improve the provision of AP in children.

**Author Contributions:** H.M. conceptualized the study. K.P. and R.B. collected and analyzed the data. K.P. and V.V. prepared the manuscript and V.V., M.H., H.M. and R.B. reviewed the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Case Report* **Suspected Malignant Hyperthermia and the Application of a Multidisciplinary Response**

#### **Laura Ebbitt <sup>1</sup> , Eric Johnson 1,\* , Brooke Herndon <sup>1</sup> , Kristina Karrick <sup>1</sup> and Aric Johnson <sup>2</sup>**


Received: 31 July 2020; Accepted: 4 September 2020; Published: 9 September 2020

**Abstract:** Purpose: Malignant hyperthermia (MH) is a critical and potentially life-threatening emergency associated with inhaled anesthetic and depolarizing neuromuscular blocker administration. This is a single center's response to MH. Summary: When signs of MH are observed, a page for "anesthesia STAT-MH crisis" is called, triggering a multidisciplinary response, including the deployment of a Malignant Hyperthermia Cart. The MH cart and the delegation of duties allows nurses, physicians and pharmacists to quickly understand their role in the stabilization, transition and recovery of a suspected MH patient. Conclusion: This case highlights the importance of multi-disciplinary involvement in these rare, but potentially fatal, cases.

**Keywords:** malignant hyperthermia; collaborative practice; perioperative care

#### **1. Introduction**

Malignant hyperthermia (MH) is a critical and potentially life-threatening emergency associated with the administration of volatile anesthetics and depolarizing neuromuscular blockers that may occur intraoperatively, as well as during the postoperative period [1]. It is treated with dantrolene, a ryanodine receptor antagonist. Both the Malignant Hyperthermia Association of the United States (MHAUS) and American Society of Anesthesiologists (ASA) emphasize a preemptive approach to treatment, including MH supply, a medication cart and departmental training [2,3]. Furthermore, delays between the onset of MH and a coordinated response involving the administration of dantrolene have been associated with increased rates of complications [1]. Therefore, a rapid and efficient response to those with suspected MH may limit the morbidity associated with the condition. We present a case of suspected MH and illustrate the application of a multidisciplinary response in accordance with a well-rehearsed institutional protocol.

## *Pathophysiology*

Malignant hyperthermia is an autosomal-dominant, pharmacogenetic disorder that manifests as a hypermetabolic crisis following exposure to a triggering agent. Known triggering agents include all volatile anesthetics (isoflurane, sevoflurane and desflurane), depolarizing neuromuscular blocking agents (succinylcholine) and human stressors such as vigorous exercise and heat [4]. The most common genetic mutation found to cause MH involves changes to the type 1 ryanodine receptor (RYR1), which encodes for the ryanodine receptor found on skeletal muscle [5]. The RYR1 is located on the sarcoplasmic reticulum of myocytes and is essential for regulating muscle excitation–contraction coupling. In the setting of genetic mutation and a triggering agent, rapid and uncontrolled increases in myoplasmic calcium occur, although this may not occur in the patients' initial surgeries. This is

significant, as both metabolism and contraction in skeletal muscle are regulated by the concentration of intracellular calcium [6]. Manifestations of the dysregulation are indicative of a hypermetabolic state. These derangements may occur as early or late signs. Early signs may include sudden elevated end-tidal carbon dioxide, tachycardia, acidosis and muscle rigidity. Late signs may include hyperthermia and hyperkalemia [4]. If untreated, these symptoms may progress to rhabdomyolysis, myoglobinuria and acute renal failure. Life-threatening complications include disseminated intravascular coagulopathy (DIC), congestive heart failure, bowel ischemia and compartment syndrome [4]. The prompt diagnosis and treatment of MH is key to preventing the progression of symptoms and avoiding significant morbidity or death.

#### **2. Institutional Approach**/**Protocol**

Prior to the administration of any anesthetic, all patients should be screened for MH through a complete medical and family history analysis. This may not be possible in emergency situations. The initial signs of MH may occur at any time following the administration of a triggering agent, including immediately following the induction of general anesthesia or at any point during the maintenance phase for the anesthetic. As previously mentioned, the earliest clinical signs include an increase in the end-tidal carbon dioxide and tachycardia. As these findings are much more frequently a result of inadequate anesthesia and hypoventilation, respectively, the anesthesiologist must maintain a high level of suspicion for MH. If the anesthesiologist feels that MH is probable, or if there is no alternative diagnosis to explain the patient's clinical findings, they should immediately discontinue any triggering agents, notify the surgeon, hyperventilate with 100% inspired oxygen, increase fresh gas flow to >10 L/min, and trigger our multidisciplinary response. If available, charcoal filters should also be placed on the inspiratory and expiratory limbs of the anesthesia circuit. As MH is a potentially lethal disorder, a well-coordinated multidisciplinary approach is valuable in ensuring a timely and organized response. Figure 1 demonstrates the sequence of events initiated at our institution when MH is suspected.

An "anesthesia stat-MH crisis" is called out over the intercom to alert operating room (OR) staff including anesthesiologists, nurses and pharmacists to respond and assist in treating the patient. The anesthesiologist will serve as the primary leader for the resuscitation response and ensure that all aspects of patient care are accounted for. The primary OR nurse will retrieve the MH cart (contents shown in Table 1) from the adjoining storage area and bring it into the OR. Color-coded cards corresponding to tasks or roles are assigned to responding personnel. These roles include a registered nurse (RN) circulator, cooling nurse, medication nurse, dantrolene nurse/pharmacist and crisis management nurse. Attached to each card is a bag of supplies specific to the individual's role. The RN circulator may assign additional MH Team roles as needed. The cooling nurse procures ice and is prepared to implement advanced cooling as indicated. Cooling techniques at our institution include ice bags at the groin, axilla and neck; cooling blankets; and cold saline, as indicated. The medication nurse starts a large bore IV and works with the pharmacist to calculate the appropriate dantrolene dose. The pharmacist double-checks all drug dosing and assists with medication documentation, as well as ensuring the order of dantrolene products are utilized in the correct order to maximize efficiency and cost-effectiveness. Additionally, the pharmacists help to procure regular insulin and dextrose if needed for the treatment of hyperkalemia. Without all of these providers assessing and participating in the care of the patient, these cases would be extremely laborious. Having a multidisciplinary team attend to an MH crisis allows for the rapid control of a patient's symptoms and to potentially stabilize them quickly.

**Figure 1.** Sequence of events involved in Malignant Hyperthermia Response.



Two types of dantrolene are contained in our MH cart, one vial of Ryanodex and nine vials of Revonto, in addition to the 10 vials of nonbacteriostatic sterile water (nine 100 mL vials and one 20 mL vial). The Ryanodex is used for the first dose, and the nine vials of Revonto are provided for any necessary subsequent dosing. Ryanodex is a lyophilized powder form of dantrolene containing 250 mg per vial, which costs around USD 2500.00. Revonto is also a lyophilized powder but in contrast only contains 20 mg of dantrolene per vial, costing around USD 60.00 a vial. To reconstitute Ryanodex, only 5 mL of sterile water is required. When reconstituting Revonto, 60 mL is needed per vial. Ryanodex should be used for the first dose because of its ease of use and need to reconstitute fewer vials. For an average 80 kg patient, 10 vials of Revonto and 600 mL of sterile water would be required to reconstitute an initial dose. Other components of the MH cart are listed below in Table 1 and follow MHAUS recommendations [7].

Once the patient's MH symptoms and the patient are clinically stabilized, post-operative critical care and intensive care unit (ICU) admission are initiated. Patient allergies are updated to include likely triggering agents as a placeholder for future operations and hospital visits as a safety measure. When appropriate, patient and family are counseled on the importance of notifying anesthesia providers about MH history and avoiding triggering agents.

#### **3. Case**

With the consent of the patient, we present a case of a 22-year-old male admitted with an open right intercondylar fracture of the distal humerus after getting his arm caught in a steel press. In the emergency department, the patient received intravenous cefazolin, morphine, hydromorphone and a Tetanus/Diptheria/Pertussis (Tdap) vaccine. He was taken to the OR on the same day for an irrigation and debridement, as well as closed reduction of the open distal humerus fracture. General anesthesia was induced with lidocaine, fentanyl and propofol. Rocuronium was used for neuromuscular blockade. During the procedure, he was maintained on sevoflurane. No complications were noted during or after initial surgery. The following day, the patient was scheduled for a definitive internal fixation of his distal humerus fracture. General anesthesia was again induced with lidocaine, fentanyl and proprofol. Succinylcholine was administered to facilitate endotracheal intubation. The patient was maintained on isoflurane, and intermittent dosing of rocuronium was used to facilitate neuromuscular blockade. Approximately 30 min into the procedure, during the placement of an additional intravenous line, unexpected resistance was noted. Upon closer examination, his extremities were found to be rigid. A quick assessment of his vitals showed that he was tachycardic, with a heart rate of 160 bpm; hypercapnic, with an end-tidal CO<sup>2</sup> of 62 mmHg; and hypertensive, with systolic blood pressures >160 mmHg (baseline blood pressure was 130/82 mmHg at preoperative evaluation). A temperature-sensing catheter was placed in the bladder, and the patient was found to be normothermic at 37.4 ◦C. Despite the normothermia, malignant hyperthermia was suspected. The isoflurane was discontinued, and charcoal filters were placed in the circuit. Nitrous oxide was used to maintain general anesthesia, and a malignant hyperthermia response was initiated and allowed for additional responders to arrive at the patient's bedside within minutes.

The patient quickly received an initial bolus of 187.5 mg of Ryanodex (2.5 mg/kg). Additionally, 20 mg of IV push esmolol was administered to treat his tachycardia but with a negligible response. Over the next 35 min, the patient received 80 mg of Revonto via intermittent 20 mg doses. These doses were administered to treat persistent and intermittent symptoms of MH.

The non-pharmacological measures taken include ice packs applied to the axilla and the placement of cooling blankets. The patient responded to the dantrolene with marked reductions in heart rate, muscle rigidity and end-tidal carbon dioxide (EtCO2). While not elevated, the patient's temperature remained normothermic. The surgical procedure was aborted, and the patient was transferred to the ICU for close monitoring, with care being assumed by the ICU intensivists.

In the ICU, the patient continued to receive Revonto 80 mg (~1 mg/kg) IV Q 6 h for 24 h. During this time period, the patient's lactate fell from 3.2 mmol/L at its peak to 0.6 mmol/L (Figure 2). His creatinine kinase (CK) peaked at 16,505 units/L and decreased to 7887 units/L prior to discharge (Figure 3). The patient's serum creatinine (SCr) was also elevated at 1.44 mg/dL and trended back down to his assumed baseline.

**Figure 2.** Arterial lactate concentration (POD = post-operative day).

**Figure 3.** Creatinine kinase concentration (POD = post-operative day).

In light of the etiology and triggering factor in this case, one may incorrectly assume it was precipitated by succinylcholine alone, since the patient previously received sevoflurane without incident. However, the literature suggests that different inhaled anesthetics may trigger MH at different rates, and his initial sevoflurane exposure was not sufficient [8]. Furthermore, studies have shown that a triggering inhalation agent plus the use of succinylcholine may cause a more marked response than a single agent [9].

After the patient was stabilized, the case was discussed with the mother, who had also experienced MH in the past; however, she was not aware that this was a hereditary disease. The patient's family was educated regarding the risks of MH and the potential for genetic predisposition within the family. An allergy was also added to the patient's chart for future potential cases. The patient was extubated that evening. Four days into the patient's admission, he received an open reduction internal fixation (ORIF) of his distal humerus. Total IV anesthesia (TIVA) was used with continuous infusion of propofol and intermittent dosing of fentanyl, dexmedetomidine and rocuronium throughout the case. Aside from the CK, lactate and SCr, the patient's lab results all remained normal, and the patient progressed to his baseline function. The patient was discharged home on post-operative day 3 from the index surgery, with follow up after 2 weeks with the orthopedic service. Through the utilization of the institution's protocol, all providers were aware of their roles within the team and were able to quickly perform their assigned duties. This allowed delays to be reduced for the rapid control of the patient's MH. Without the swift initiation of an MH protocol, it is possible that patients could experience a lethal outcome.

#### **4. Conclusions**

MH is a rare but serious metabolic complication associated with the use of volatile anesthetics and depolarizing neuromuscular blocking agents. In the case of a delayed response or missed diagnosis, significant morbidity and mortality may occur. Institutions should develop, implement and train staff on how to recognize and treat this acute disorder. We present the case of a patient with an unknown family history of malignant hyperthermia. Despite proper pre-operative assessment, the family history was missed, and the patient experienced MH symptoms after receiving a triggering agent during his second surgery. Due to an extensive, multidisciplinary perioperative MH protocol, this patient was successfully treated and avoided serious complications. Providers were able to treat the patient quickly and efficiently, in great part due to the presence and utilization of the MH cart. The dosing cards and instructions readily available on the cart allowed the correct dose of Ryanodex to be verified and drawn up into a syringe by the providers while subsequent doses of Revonto were also being prepared. This case also highlights the need to ask specific questions in the pre-operative setting regarding both the patient's and the patient's family's prior history of surgeries and any events that may have occurred. We recommend that other institutions develop a similar cart, as a mechanism for providers to be able to respond to these events.

**Funding:** This research received no external funding.

**Conflicts of Interest:** All authors report no conflicts of interest, including pharmaceutical or industry support, regarding any of the information contained in this report. No relevant funding from any organization was provided to any of the authors regarding this manuscript or the ideas contained herein.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Remote Monitoring of Critically-Ill Post-Surgical Patients: Lessons from a Biosensor Implementation Trial**

**Mariana Restrepo <sup>1</sup> , Ann Marie Huffenberger <sup>2</sup> , C William Hanson III 2,3, Michael Draugelis <sup>4</sup> and Krzysztof Laudanski 5,6,7,\***


**Abstract:** Biosensors represent one of the numerous promising technologies envisioned to extend healthcare delivery. In perioperative care, the healthcare delivery system can use biosensors to remotely supervise patients who would otherwise be admitted to a hospital. This novel technology has gained a foothold in healthcare with significant acceleration due to the COVID-19 pandemic. However, few studies have attempted to narrate, or systematically analyze, the process of their implementation. We performed an observational study of biosensor implementation. The data accuracy provided by the commercially available biosensors was compared to those offered by standard clinical monitoring on patients admitted to the intensive care unit/perioperative unit. Surveys were also conducted to examine the acceptance of technology by patients and medical staff. We demonstrated a significant difference in vital signs between sensors and standard monitoring which was very dependent on the measured variables. Sensors seemed to integrate into the workflow relatively quickly, with almost no reported problems. The acceptance of the biosensors was high by patients and slightly less by nurses directly involved in the patients' care. The staff forecast a broad implementation of biosensors in approximately three to five years, yet are eager to learn more about them. Reliability considerations proved particularly troublesome in our implementation trial. Careful evaluation of sensor readiness is most likely necessary prior to system-wide implementation by each hospital to assess for data accuracy and acceptance by the staff.

**Keywords:** wearable biosensors; critical care; vital sign monitoring; bio-monitoring system; technology acceptance; integration; implementation

#### **1. Introduction**

The ability of biosensors to wirelessly, un-obstructively, and effortlessly monitor patients has become a fascinating prospect for healthcare [1]. They offer an opportunity to improve patient care while reducing costs and increasing patient and staff satisfaction [2,3]. At a minimum, most biosensors collect body temperature, pulse, heart rate variability, respiration rate, peripheral capillary oxygen saturation (SpO2), sleep, and movement. Although sensors can quite often deliver additional data, it is unclear if they can increase the effectiveness of healthcare delivery.

**Citation:** Restrepo, M.; Huffenberger, A.M.; Hanson, CW., III; Draugelis, M.; Laudanski, K. Remote Monitoring of Critically-Ill Post-Surgical Patients: Lessons from a Biosensor Implementation Trial. *Healthcare* **2021**, *9*, 343. https://doi.org/10.3390/ healthcare9030343

Academic Editor: Richard H. Parrish II

Received: 13 February 2021 Accepted: 6 March 2021 Published: 18 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

In order to effectively integrate biosensors into healthcare workflow, several factors have to be fulfilled [4]. Foremost, the reliability of the equipment needs to be assessed. A previous study found that when comparing SpO<sup>2</sup> measurements between five types of biosensors and a clinical vital sign monitor, a range of 85–100% of biosensor measurements fell within three percentage points of the clinical monitor, depending on the type of biosensor [5,6]. However, the same study alternatively established that this range shifted to 93.5–100% of biosensor measurements falling within three beats per minute (BPM) of the clinical monitor [5]. It is also notable that mean skin temperature measured by biosensors can vary up to 2 ◦C from axillary measurements [7]. Furthermore, recordings from the research-grade biosensors proved less accurate than those intended for consumers [6,8,9]. Both the consistency and accuracy of some vital signs are much dependent on the device model [6,8,9]. Finally, the devices must take into account features specific to patients [10]. These inconsistencies across differing vital signs could introduce deceptive data trends that would undermine the feasibility of implementing biosensors in a critical care setting.

The implementation of biosensors in the workflow must be very well-planned and unit-specific [11]. The demands for perioperative care are particularly sensitive to interruption of the signal, while in other instances, accuracy may matter more. The data has to be delivered from sensors via a secure wireless network connection to provide a clear advantage over the existing infrastructure [4]. Establishing such a link securely and reliably is a complex task, especially in a hospital system with multiple entities operating off varying information system infrastructures [12]. Providing similar monitoring at home is even more complex. Acceptance of the sensor must be high across all parties involved [4,11,13,14]. Patients should value the sensor as an improvement over prior solutions. Sensors should be especially comfortable and undisruptive in perioperative settings. Providers should expect robust and reliable sets of data adding to the care being provided. Similarly, nursing staffs seek to ease the burden of continuously monitoring patients remotely, allowing biosensors to improve the quality and safety of patient care. All these requirements are particularly important for perioperative care, especially in in-home settings. A useful framework for the implementation of biosensors is provided by the ABCDEF bundle by suggesting a focus on which parameters yield most of the value [15]. Understanding potential barriers to this integration is the key to major transformations in healthcare [4,9,10,16].

This study describes the process of implementation of a multisensory biosensor platform to analyze up to 22 parameters and features in intensive care unit (ICU) patients. We aimed to describe our implementation process experiences, with special emphasis on comparing data streams from patients being monitored by biosensors versus standard hospital physiological monitoring. We also analyzed acceptance of the technology by patients, providers, and nurses. Past studies have found that while biosensors have extensive potential for real-world adaptations, functional challenges, including data validity and stability, need to be overcome first before defining practical applications [17].

#### **2. Materials and Methods**

The IRB at the University of Pennsylvania approved the study (#832633). Data were collected in 2020.

This is a pilot study testing the feasibility and robustness of the two types of wearable biosensors in anticipation of future deployment. One of the sensors is commercially available and used predominately for personal care, and it has not been previously tested in a healthcare ICU setting. The other one represented a biosensor that was developed and manufactured for healthcare use by a start-up. Both sensors collect several parameters, but we only focus on the data which are collected by the standard for medical ICU monitoring (Nihon Kohden USA; Irvine, CA, USA). The vital signs this study focuses on include heart rate, respiratory rate, and peripheral capillary oxygen saturation, as these can be collected by both types of biosensors and the Nihon Kohden monitoring system.

The study was conducted in an eight-bed medical ICU. The staff consists of an attending pulmonologist, one advanced practice provider, and four to five nurses. They were

introduced to the study and hardware during a brief 10-min orientation. Patients were approached for consent while being in the ICU. Seven individuals agreed to participate, while one refused. One individual wore two sensors subsequently. The demographic characteristics of the study subjects are detailed in Table 1. After consenting, a patient was fitted with a sensor using the respective manufacturer's recommendation. The staff was instructed to keep the sensor on for a 24-h period. After the collection of data, the sensor was removed. Patients and staff members were asked to complete a quick survey in the RedCap database (Appendix A.1) [18]. In addition, we asked the staff to complete a separate survey after the trial period to explore their perception of biosensors (Appendix A.2).


**Table 1.** Demographic characteristics of studied cohorts.

The data obtained from the biosensors were analyzed and compared to standard clinical monitoring provided using correlation and pathway analysis. Parametric variables were expressed as mean ± SD and compared using a Student's *t*-test. For non-parametric variables, median (Me) and interquartile ranges (IR) were computed. Mann–Whitney U statistics were employed to compare non-parametric variables. Data groups were analyzed as independent groups. A double-sided p-value of less than 0.05 was considered statistically significant for all tests. The r-Pearson statistic was calculated to determine the correlation between the studied variables. Statistical analyses were performed using Statistica 11.0 (StatSoft Inc., Tulsa, OK, USA). Graphs were generated using GraphPad Prism 8.4.2 (GraphPad Software Inc., San Diego, CA, USA).

#### **3. Results**

#### *3.1. Data Accuracy*

The biosensors' data showed varied performances with respect to different vital signs. Compared to respiratory rate and peripheral capillary oxygen saturation (SpO2), heart rate measurements demonstrated the strongest and most consistent correlation between a biosensor and wired ICU standard recordings at rest (Figure 1A(i),(iii)) and during movement (data not shown). Although the quality of the heart rate data fluctuated throughout this specific trial, it remained above 80% for most of the measurements recorded after the application of the biosensor (Figure 1A(ii)). The difference between the biosensor's recordings and those of the Nihon Kohden system is assessed as the bias of the measurements, which is minimal and optimal for heart rate readings (Figure 1A(iv)). However, one trial demonstrated a significant lapse in the correlation during the onset of the measurements. The quality of the biosensor's measurements during this time was significantly less than once the heart rate stabilized.

**Figure 1.** Correlation between data supplanted by multimodal sensor and standard ICU monitoring. Various degrees of data consistency were demonstrated by biosensors ranging from excellent for heart rate measurements (**A**), to variable for respiratory rate observations (**B**), to suboptimal SpO<sup>2</sup> recordings (**C**). In addition to the vital signs measured (**i**) and the quality of the biosensor measurements (**ii**), the correlation (**iii**), and bias (**iv**) between biosensor and Nihon Kohden recordings were also reported according to vital sign.

The correlation between biosensor and monitor-driven measurements for respiratory rate was significantly more variable than that of the heart rate recordings. One sample displayed superficially close correlations with similar results for both the biosensors and the manual measurements (Figure 1B(i)). This was confirmed by the weak positive relationship seen on the scatter plot that described the correlation between the two types of measurements (Figure 1B(iii)). Similar to the heart rate sample previously discussed, the quality of the measurements fluctuated throughout the trial, especially in the first half (Figure 1B(ii)). The bias reporting the difference between the biosensor and Nihon Kohden respiratory rates is visibly more than that seen for the heart rate data, further emphasizing the increased variability between the two forms of recording (Figure 1B(iv)).

The SpO<sup>2</sup> measurements showed the most variability in terms of the correlation between the biosensor and standard monitor measurements. Most samples reflected no SpO<sup>2</sup> measurements on the biosensors' parts (Figure 1C(i)). This lack of recording was seen in at least three different samples. Interestingly, the evaluation of the biosensors' quality did not reflect this, and instead remained at above 80% for the majority of the trial (Figure 1C(ii)). On another occasion, the biosensor only recorded periodically and at various qualities (data not shown). Similarly, the corresponding scatterplot for this sample does not reflect any correlation between the types of measurements (Figure 1C(iii)). The difference between the biosensor and standard monitor recordings seems to be greater than that of the heart rate measurements, as supported by the bias diagram (Figure 1C(iv)).

#### *3.2. Deployment of the Sensors*

The perspectives of patients wearing the biosensors, providers wearing the biosensors (providers as subject), and providers applying the biosensors on patients (provider for patients) were obtained through questionnaires to gauge the operationalization and ease of implementation of the biosensors. Determining the form factor and acceptance related to the biosensors is critical because these factors drive the discussion on implementation using the perspectives of both patients and providers. Specifically assessing the viewpoint of providers wearing the biosensors serves as an interesting comparison in relation to that of the patients they are treating.

The devices' adherence to the skin was perceived as somewhat problematic by healthy individuals. Despite small form factor, most of the users and medical staff considered sensors to interfere with daily activities (Figure 2). Medical staff included MDs (medical doctors) and RN (registered nurses). Irritation was reported by a minority of the patients, with one individual reporting skin abrasion out of a total of eight patient trials (Figure 2). Only one trial was terminated before the prescheduled time because of the irritation. The operationalization of the sensor was assessed very highly by patients wearing them when asked how much they agreed with the following: "Did you like the way the biosensor fit?", "Was it easy to apply?", "Was it easy to connect?", "Was it easy to remove?", and "How was your overall experience related to biosensor?" (Figure 2). Finally, the sensor trials were terminated on time, at the prescheduled time, in all study groups (providers as subjects = 65%, patients = 75%, providers for patients = 92%). Neither of the clinical groups discontinued the sensor because of interference with clinical care.

#### *3.3. Perception of the Sensors*

There was little difference in perception of the different domains of the sensors' usability between MDs and RNs, except for the familiarity with sensors between RNs and MDs (Figure 3A). The most common positive comments about sensors were "modern/sleek", "mobility", and "more data". The most common negative adjectives were "application", "unreliable", and "cost". The major sensor advantages were "easy application", "notobstructive", and "portability".

The majority of MDs and RNs believed that sensors would be deployed in the next 3 to 5 years (B). The staff was feeling relatively unprepared for sensor deployment (Figure 3C).

**Figure 2.** Experience of wearing the sensor. Experience of wearing the sensor was consistently rated higher for patient users compared to providers involved in care of patients.

**Figure 3.** Readiness for implementation of biosensors. Physicians assessed the benefits of sensor deployment highly (**A**) and predicted faster implementation (**B**) than nurses. Nurses reported a more slightly unprepared perception of readiness to work with biosensors (**C**).

#### **4. Discussion**

The implementation of biosensors demonstrated several important related problems. The reliability of a sensor has to be extensively studied before the implementation. Prior reports pointed to unique problems related to the biosensors, although this was not the uniform case [12,16,19,20]. Movement, skin color, and sweating were quite often reported

as the main reasons for interference [12,19]. Post-deployment interviews demonstrated that data might be lost for other reasons [14]. Sensor adherence was cited as such, but the loss of some data could not be explained exclusively. Considering that the correlations between the biosensor and clinical recordings for respiratory rate and SpO<sup>2</sup> were not significantly accurate, the variability that was introduced could negate the reliability and accuracy of the biosensors [21–24].

Overall, data correlation depended more on the data type (e.g., vital sign recorded) than on the sensor type in our study, and that was a new finding [12,16,19]. The weak correlations between readouts of the sensor and clinical standards augment the skepticism regarding integrating the biosensors with more standard critical care technology. Without a standard for accuracy, the variability will require consistent validation of the results, which will be both time sensitive and concerning if the validation fails. These problems emerged even before we could test the sensors' connection to the IT system. The unpredictability of the biosensors connecting to the appropriate downloading devices or tablets is one of the main concerns regarding this novel technology [4]. Being unable to anticipate if or where the biosensor will connect is one possible restriction that diminishes the fidelity of biosensors, given they should function to wirelessly monitor patients at all times. The stability and resiliency, among other technological obstacles, of electrochemical biosensors have proven to be focal points for barriers to their implementation, and the acceptance rate for loss signal has not being established [10]. However, our study demonstrated that multi-sensor devices might be uniquely prone to sensing errors as compared to clinical standards. This is a new and unique finding [12,16].

The adverse effects of wearing the sensors were rare. Irritation was almost not observed, while only one case of abrasion was noted in our study. The small number of enrolled subjects precluded this from being a conclusive study. Future studies should look into the incidence of adverse effects related to biosensors' application as compared to regular monitoring. However, most of the devices are fairly inert while being worn by patients [6,10,12,16].

The acceptance of the biosensor technology was particularly high for patients and slightly less so among the providers. This was the novel finding of the study, since some reported several barriers [14,16]. The reason driving the high acceptance of the biosensors was the relatively low form factor of devices [4,21,23]. A desire for non-interference of the device was frequently cited [16]. We demonstrated relatively low initial enthusiasm at the beginning of the trial that significantly increased at the completion of trials. Patients had overall positive impressions. The interference with workflow was minimal, though providers wearing the sensors reported much higher rates of premature termination of the trials secondary to adherence problems. The increased mobility of healthy individuals compared to bedridden patients may be partially responsible for this difference [14,16].

Our study has several limitations. This was not a device trial, or even a pilot study. The sample size was small, and we used two different devices. Devices were placed on few patients or staff members. However, the intention of this paper was to observe the implementation process to demonstrate potential problems. Much too often, the problems during implementation are not brought up, setting unrealistic expectations from the enduser.

#### **5. Conclusions**

We caution against an overoptimistic approach to the implementation of biosensors in a healthcare setting, as the process has several potential pitfalls. Despite being FDAapproved, biosensors need to be consistently tested against standard monitoring equipment, such as that of Nihon Kohden, in order to demonstrate readiness for implementation in high-acuity healthcare settings.

**Author Contributions:** M.R.—data analysis, manuscript writing, A.M.H.—design and supervision, CW.H.III—supervision, manuscript writing, M.D.—data analysis, reliability, K.L.—concept, design, data collection, analysis, manuscript writing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

**Institutional Review Board Statement:** The study was approved by the Institutional Review Board at the University of Pennsylvania.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request.

**Acknowledgments:** We would like to thank Sean Sarles for his able assistance.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

*Appendix A.1. The Questionnaire Used to Study the Attitude of the Staff towards Biosensors in Patients and Staff*

	- a. Attending
	- b. APP
	- c. RN
	- d. CNA
	- e. Other staff
	- a. Y/N

a. 0–10

	- a. Mobility (0—not at all; 10—extremely)
	- b. sleep (0—not at all; 10—extremely)
	- c. transport? (0—not at all; 10—extremely)
	- a. How esay is the biosensor to use (0—not at all; 10—very)
	- b. How practical Is the biosensor to use in the healthcare setting (0—not at all; 10—very)
	- a. Never
	- b. 1–2 years
	- c. 5–6 years
	- d. In 10 years
	- e. Never
	- a. X
	- b. X
	- c. X

*Appendix A.2. The Questionnaire Used to Study the Attitude of the Staff towards Biosensors in Patients and Staff*

	- a. Attending
	- b. APP
	- c. RN
	- d. CNA
	- e. Other staff
	- a. Y/N
	- a. 0–10
	- a. Never
	- b. 1–2 years
	- c. 5–6 years
	- d. In 10 years
	- e. Never
	- a. X
	- b. X
	- c. X

#### **References**


## **A Meta-Analysis on Prophylactic Donor Heart Tricuspid Annuloplasty in Orthotopic Heart Transplantation: High Hopes from a Small Intervention**

**Alberto Emanuel Bacusca 1,2,†, Andrei Tarus 1,2,†, Alexandru Burlacu 2,3,\* , Mihail Enache 1,2 and Grigore Tinica 1,2**


**Abstract:** (1) Background: Tricuspid regurgitation (TR) is the most frequent valvulopathy in heart transplant recipients (HTX). We aimed to assess the influence of prophylactic donor heart tricuspid annuloplasty (TA) in orthotopic HTX (HTX-A), comparing the outcomes with those of HTX patients. (2) Methods: Electronic databases of PubMed, EMBASE, and SCOPUS were searched. The endpoints were as follows: the overall rate of postprocedural TR (immediate, one week, six months, and one year after the procedure), postoperative complications (permanent pacemaker implantation rate, bleeding), redo surgery for TR, and mortality. (3) Results: This meta-analysis included seven studies. Immediate postprocedural, one-week, six-month and one-year tricuspid insufficiency rates were significantly lower in the HTX-A group. There was no difference in permanent pacemaker implantation rate between the groups. The incidence of postoperative bleeding was similar in both arms. The rate of redo surgery for severe TR was reported only by two authors. In both publications, the total number of events was higher in the HTX cohort, meanwhile pooled effect analysis showed no difference among the intervention and control groups. Mortality at one year was similar in both arms. (4) Conclusion: Our study showed that donor heart TA reduces TR incidence in the first year after orthotopic heart transplantation without increasing the surgical complexity. This is a potentially important issue, given the demand for heart transplants and the need to optimize outcomes when this resource is scarce.

**Keywords:** heart transplant; tricuspid annuloplasty; tricuspid regurgitation; prophylactic; meta-analysis

### **1. Introduction**

Tricuspid regurgitation (TR) is the most frequent valvulopathy in heart transplant recipients (HTX), with a reported incidence ranging between 19% to 84% [1,2]. The tricuspid valve (TV) integrity manifests a significant impact on the long-term clinical progress and survival of orthotopic HTX. Although most of the patients present a small degree of tricuspid insufficiency, moderate or greater grades were associated with significantly worse survival and higher post-transplant complications [3]. TR etiology is multifactorial, with several viable hypotheses still debatable: biatrial transplantation technique, allograft dysfunction or rejection, donor-recipient size mismatch, or structural damage during endomyocardial biopsy [4–8].

Postoperative moderate or severe TR negatively affects the overall survival rates after HTX [9]. Despite the fact there is a reported improvement of the degree of tricuspid regurgitation six months after the transplantation, the nature of this valvulopathy is

**Citation:** Bacusca, A.E.; Tarus, A.; Burlacu, A.; Enache, M.; Tinica, G. A Meta-Analysis on Prophylactic Donor Heart Tricuspid Annuloplasty in Orthotopic Heart Transplantation: High Hopes from a Small Intervention. *Healthcare* **2021**, *9*, 306. https:// doi.org/10.3390/healthcare9030306

Academic Editor: Richard H. Parrish II

Received: 18 February 2021 Accepted: 8 March 2021 Published: 10 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

progressive. Studies with more extended follow-up periods reported an increase in severe TR incidence from 7.8% at five years to 14.2% at ten years [10].

The most frequently reported indication for heart surgery after HTX was the atrioventricular valve reconstructions or replacement. 62.5% of these cases were related to the tricuspid valve [11]. Surgical repair or replacement is required when right heart failure becomes refractory to conservative medical treatment [10,12]. The mean duration from transplantation to severe TR diagnosis is reported to be 43 +/- 6.38 months [10]. The cardiac mechanics portending right ventricular failure can be accurately predicted using either right cardiac catheterization or by noninvasive methods computational modeling of hemodynamic and cardiac mechanics using lumped-parameter and biventricular finite element analysis [13,14].

To improve the TV function and avoid the risks associated with redo heart surgery, prophylactic tricuspid annuloplasty (TA) on the donor's heart was proposed as a simple solution to a problem that triggered an increasing concern. Already an established and widely performed surgery, primarily in functional TR treatment, TA accomplished either by DeVega's technique or by a ring is associated with excellent long-term results [15,16]. TA was envisioned to enhance posttransplant hemodynamics and prevent late moderate/severe TR. Moreover, the importance of TV repair was emphasized not only in heart transplanted patients but also in those receiving left ventricular assist devices either as a bridge therapy or as destination therapy, in which concomitant TV repair may reduce postoperative right ventricular failure [17].

Although a significant reduction in TR after this procedure was reported by most of the authors, actual data are controversial, and opinions regarding its impact on overall survival are heterogeneous. To date, there is no consensus on the concomitant management of the TV during heart transplant [18].

The purpose of this study is to assess the influence of prophylactic donor heart tricuspid annuloplasty (in terms of postoperative complications, effects on hemodynamic parameters, short- and long-term tricuspid regurgitation, and mortality) in orthotopic heart transplant recipients.

#### **2. Materials and Methods**

The preferred reporting items for systematic reviews and meta-analysis (PRISMA) checklist was applied in each step of the meta-analysis conduction (Supplementary Table S1).

#### *2.1. Search and Eligibility*

We performed an extensive search for studies comparing heart transplantation with and without prophylactic tricuspid annuloplasty in three electronic databases: PubMed, EMBASE, and SCOPUS from inception to 20th December 2020. We used the following interrogation terms: "heart transplantation," "tricuspid regurgitation," "tricuspid valvuloplasty," "de Vega." Two independent authors (A.E.B. and A.T.) checked titles and abstracts for eligibility. Fulltext was retrieved for selected papers and verified for fulfilling the following inclusion criteria: (1) study design—randomized control trials, observational studies, propensity score match studies; (2) population—patients with orthotopic heart transplantation; (3) intervention—donor heart tricuspid annuloplasty; (4) comparators—heart transplanted patients without prophylactic tricuspid annuloplasty; (5) outcomes—reported at least post-transplantation tricuspid regurgitation. Both authors scanned the references in relevant articles. The third reviewer (G.T.) mediated the situations when consensus regarding a manuscript's inclusion was not achieved.

#### *2.2. Intraoperative Timing and Outcomes*

We compared intraoperative timing between two cohorts (ischemic time, cardiopulmonary bypass time, and cross-clamp time). The endpoints were as follows: the overall rate of postprocedural TR (immediate, one week, six months, and one year after the procedure), postoperative complications (permanent pacemaker implantation rate, bleeding), redo surgery for TR, and mortality.

#### *2.3. Data Collection and Synthesis*

The same reviewers extracted data only from retrieved published manuscripts and registered them in standard tables. When the ratio of events and not raw data were available, we calculated the event number from the described ratio and total cohort.

Review Manager (RevMan) Version 5.3 (Nordic Cochrane Centre, The Cochrane Collaboration, 2012, Copenhagen, Denmark) software was used to generate the pooled effect size with odds ratio (OR) and 95% confidence intervals (CI) by Mantel–Haenszel method and random effect model for dichotomous data. A *p*-value of less than 0.05 was considered significant. Conversion to mean and standard deviation (SD), when median and IQR were available, was performed following the methods published by Luo et al. and Wan et al. [10,11]. The pooled sample mean and pooled standard deviation for selected studies were calculated according to the Cochrane Handbook's recommendation for Systematic Reviews. We used MedCalc Statistical Software version 14.8.1 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org (accessed on 20 December 2020); 2014) for comparative statistics. Chi-squared and *t*-Student's tests were used to compare dichotomous and continuous data.

#### *2.4. Studies Quality Assessment*

The risk of publication bias was assessed with the Newcastle–Ottawa quality assessment scale (NOS) for cohort studies and the Cochrane risk of bias tool for randomized controlled trials.

#### **3. Results**

#### *3.1. Literature Search and Study Selection*

The digital search identified a total of 1506 titles. After duplicates removal, a total of 1068 references were screened by title and abstract. There were 26 articles selected for full-text analysis (Figure 1).

Seven full-text articles that compared the incidence of moderate or severe tricuspid regurgitation, postoperative complications, and late mortality in heart transplant patients with donor tricuspid annuloplasty with cohorts with no prophylactic tricuspid valve repair during OHT were retrieved [2,19–24]. Two of the studies had the same cohort of patients and reported the same outcomes at different periods [22,23]. Two other studies have been conducted by the same authors in the same center [2,20]. The criteria for patient selection and the reported outcomes were the same. We have considered the data presented in the most recent study that also included the more representative cohorts of patients.

**Figure 1.** Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram for study selection.

#### *3.2. Study Characteristics and Risk of Bias*

The characteristics of the selected studies are presented in Table 1. All studies were appreciated to have a good quality design (Supplementary Table S2).

### *3.3. Patient and Periprocedural Characteristics*

The final analysis included 730 patients, of which 359 heart transplant recipients with prophylactic donor tricuspid annuloplasty (HTX-A) and 371 patients without tricuspid valve repair (HTX group). Both bicaval and biatrial heart transplantation techniques were taken into account. De Vega and Ring tricuspid valve annuloplasty procedures were analyzed.

Baseline characteristics and periprocedural data distinguishing each group are summarized in Table 2. Patients in both groups predominantly male and had similar ages.


#### **Table 1.**Summary of included studies.

115

HTX—heart transplantation; HTX-A—heart transplantation with tricuspid annuloplasty; TVA—tricuspid valve annuloplasty.

#### **Table 2.**Baseline characteristics and periprocedural data.


Ischemic etiology of the end-stage heart failure was more frequent in the HTX group (88.63% vs. 67.57%, *p* = 0.0001). There was no difference in preoperative renal status, mechanical circulatory support, or inotropic drug use. The pulmonary capillary wedge pressure was higher in the HTX-A group (19.70 ± 9.13 vs. 17.15 ± 8.54, *p* = 0.0047), but pulmonary vascular resistance was similar.

**2021**, , x 6 of 13

#### *3.4. Intraoperative Times*

Intraoperative data analysis revealed longer cardiopulmonary bypass time (173.32 ± 27.75 vs. 154.14 ± 25.88, *p* < 0.0001) and ischemic time (181.75 ± 40.82 vs. 165.31 ± 41.72, *p* < 0.0001) in the HTX group, but no difference in the aortic cross-clamp time.

#### *3.5. Outcomes*

#### 3.5.1. Tricuspid Regurgitation

Forest plots for postoperative TR in different periods are shown in Figure 2a–d. Immediate postprocedural, one week, six months and one year tricuspid insufficiency rate was significantly lower in HTX-A group (HTX-A vs. HTX: OR: 0.04, 95% CI, 0.01 to 0.34, I <sup>2</sup> = 0%); (HTX-A vs. HTX: OR: 0.25, 95% CI, 0.06 to 1.03, I<sup>2</sup> = 8%); (HTX-A vs. HTX: OR: 0.18, 95% CI, 0.05 to 0.66, I<sup>2</sup> = 0%); (HTX-A vs. HTX: OR: 0.17, 95% CI, 0.04 to 0.77, I<sup>2</sup> = 0%).





**Figure 2.** Forest plot depicting post-transplantation TR: (**a**) immediate; (**b**) after 1 week; (**c**) after 6 months; (**d**) after 1 year.

#### 3.5.2. Periprocedural Complications

There were no difference in permanent pacemaker implantation rate between the goups (HTX-A vs. HTX: OR: 2.19, 95% CI, 0.50 to 9.64, I<sup>2</sup> = 0%) (Figure 3a). Incidence of postoperative bleeding was similar in both arms (HTX-A vs. HTX: OR: 1.00, 95% CI, 0.23 to 4.28, I<sup>2</sup> = 0%) (Figure 3b).

**Figure 3.** Forest plot depicting periprocedural complications: (**a**) permanent pacemaker implantation rate; (**b**) postoperative severe bleeding rate.

3.5.3. Reoperation and Survival

The rate of redo surgery for severe TR was reported only by two authors. In both publications, the total number of events was higher in the HTX cohort, meanwhile pooled effect analysis showed no difference among the intervention and control groups (HTX-A vs. HTX: OR: 0.13, 95% CI, 0.02 to 1.11, I<sup>2</sup> = 0%) (Figure 4a). Mortality at 1 year was similar in both arms (HTX-A vs. HTX: OR: 1.01, 95% CI, 0.41 to 2.49, I<sup>2</sup> = 0%) (Figure 4b).


**Figure 4.** Forest plot depicting: (**a**) reoperation rate on tricuspid valve after transplantation; (**b**) 1-year survival.

#### **4. Discussion**

Our meta-analysis shows that donor heart tricuspid annuloplasty reduces tricuspid regurgitation incidence in the first year after orthotopic heart transplantation without increasing the surgical complexity. No significant benefit or harm was revealed on longterm mortality. Performed in high-experienced centers, prophylactic donor tricuspid annuloplasty could be routinely considered during orthotopic heart transplantation as it tends to incline the balance to a more favorable evolution.

Tricuspid regurgitation is a common problem after heart transplantation. There are two main types of tricuspid insufficiency. Type I dysfunction is more common and occurs earlier, with a reported average time from the procedure to the onset of severe TR of 13 months [25]. In this scenario, the regurgitation is due to the alteration in the TV geometry and right atrium, followed by annular/ventricular dilation. The tricuspid valve leaflet motion is normal. Evolution under medical therapy is usually mild but may become severe and require surgical correction [25,26].

Type II dysfunction has a reported average time to onset of severe TR of 28 months and is characterized by an excessive leaflet motion mostly due to chordal disruption after right ventricular endomyocardial biopsy [25]. Mild to moderate TR may be well-tolerated, but recurrent injury or spontaneous rupture of the chordae tendineae could also lead to severe symptomatic TR that may require surgical repair [27,28].

The etiology of the disease is multifactorial. In a multivariate analysis, the standard biatrial transplantation technique is considered the most independent predictor for early and late TR in heart transplant recipients [5]. Due to a higher distortion and dilatation of the tricuspid annulus, biatrial transplantation can lead to a more frequent and severe type I tricuspid regurgitation in all time scales following transplantation. After a one-year follow-up, the patients who underwent transplantation by the biatrial technique showed higher right-sided pressures and thus added another risk factor in developing the TR [5].

Despite these findings, some authors disagree with this hypothesis. Kim and colleagues found that the occurrence of TR was not related to the anastomosis technique [29], and Kalra et al. revealed in an echocardiographic study comparing bi-caval versus atrial anastomosis technique, no effect of the technique on tricuspid regurgitation [30]. Another study identified that the strongest predictor of moderate to severe TR would rather be the presence of intraoperative RV dysfunction [3]. Other risk factors associated with the development of type I TR are the donor age, the preoperative pulmonary hemodynam-

## (**a**)

ics, pre-transplant dilated cardiomyopathy weight mismatch, and more than two cellular rejection episodes [5,9,29].

The development of long-term significant type II TR after transplantation was correlated with the number of endomyocardial biopsies performed (EMB) [5]. (*A significant correlation between the occurrence of tricuspid valve injury and EMB number performed per patient was observed* [12].) Percutaneous transvenous EMB remains the most suitable method for the early identification of histopathologic alterations; thus, the gold standard in the diagnose of cardiac rejection [31]. The reported TR caused by iatrogenic injury during EMB was 6–32% of cases [3,32,33], and almost half of all myocardial fragments recovered from patients with significant TR revealing the presence of chordae tendineae [12]. The risk factors of developing tricuspid injury are EMB technique, bioptome type, method of bioptome guidance, and access route and team experience [12,32,34]. Noninvasive methods sought to replace the EMB yet did not prove able to overcome histological analysis's advantages [35,36]. Gallium-67 scintigraphy used as a screening method has resulted in favorable outcomes, with an approximately 10-fold reduction of EMB per patient [37]. Although TV annuloplasty is performed to maintain the annulus's standard size, minor structural damage caused by EMB could also be attenuated due to the annulus reduction [23].

The impact of TR on transplantation outcomes is unquestionable. Anderson and colleagues report a 38% operative mortality in patients with mild or greater severity TR versus 7% in patients with no or trace TR. In the absence of RV dysfunction, one-year survival rates were 92% for those with no or trace TR vs. 57% with mild or greater severity TR. A vital survival gap was also noticed in the patients with RV failure (83% vs. 63%) [3]. After ten years, follow-up in Algharni et al. reported 90% survival rates in patients with less than moderate tricuspid regurgitation compared to 43% for moderate and severe TR [9]. Individuals with higher grades of TR also had more extended hospital stays and higher renal dysfunction rates and dialysis [18]. They were also more prone to need mechanical circulatory support and required more often redo open chest procedures [3].

Although prompt surgical repair of severe TR that develops early after transplantation is regarded as a safe procedure in selected patients, with an improvement in the overall survival after 1, 5 and 10 years due to better cardiac performance and alleviation of associated organ dysfunction, this redo surgery is not risk-free [11,38]. The postoperative evolution was marked by high rates of prolonged ventilation (33%), new-onset requirement of hemodialysis treatment (36.8%), and infectious complications (11.1%). The reported early mortality was 11.1% [11].

Tricuspid valve annuloplasty had been proven already as a simple, safe, effective, and reliable surgical procedure [39]. Moreover, because it is the least expensive way to treat functional TR, De Vega's TVA established itself as the treatment of choice for functional TR [39]. The procedure adds little additional time of 5 to 10 min to the operation, the fact that it is also suggested by similar aortic cross-clamp times between the HTX and HTX-A groups [40]. Instead, our results show that TVA contributed to a shorter cardiopulmonary bypass and ischemic time fact attributed to improved right ventricular performance and hemodynamic parameters [23].

TVA has been hypothesized to exert its significant benefits in the early postoperative period [23]. Our meta-analysis of immediate postprocedural, one-week, six-month, and one-year tricuspid insufficiency rates showed significantly lower values in the HTX-A group. This finding would explain the rationale behind establishing the prophylactic donor tricuspid annuloplasty procedure as standard practice. On the other hand, contrary to expected, there was no other significant improvement in the postoperative outcomes. Even though multiple authors have brought strong arguments about the TR's impact on morbidity and mortality rates, our results revealed no difference in one-year mortality between groups. Unfortunately, the fact that survival data were very heterogenous reported could be why these inexplicable results.

One of the most significant drawbacks of the procedure revolves around the complications involving the conduction system. Rubin and colleagues conducted the most edifying

study that focuses on the electrophysiologic consequences associated with tricuspid annuloplasty in heart transplantation. The conduction disturbances reported as significantly more common in the experimental group were the right bundle branch block, left anterior fascicular block, and complete heart block. Permanent pacemaker (PPM) implantation was also more frequent in patients receiving DVA. The authors advise that annuloplasty should be integrated within the context of an equitable tradeoff between the possible risk of conduction abnormalities that occur in the immediate postoperative period and the benefit of preventing late moderate/severe TR [24].

The reported incidence of PPM implantation in heart transplanted patients varies between 5.3% and 10.9% [24,41,42]. Older patients undergoing a biatrial surgical technique with a previous history of amiodarone use are already more susceptible to necessitate pacing without tricuspid intervention [43,44]. Our results showed no difference in the PPM between the groups. However, the negative effect of tissue-damaging during annuloplasty may have been counterbalanced by a shorter ischemic time in the HTX-A group previously reported to contribute to the occurrence of the conduction disturbances [44].

All in one, TA is a simple technique that is worth considering when it comes to orthotopic heart transplantation. The procedure's aim is clear: to reduce the annulus dilatation development and thus the long-term tricuspid regurgitation. If the results are according to what was initially expected when they were first introduced is still debatable. Correctly performed, it could reduce the risk of severe regurgitation and thus, improve the survival rates and postoperative outcomes while carrying no additional risk for the patient. Some surgeons have discontinued this procedure two years after its implementation, some have assimilated it into the transplantation protocol on the presumption that it has its advantages. However, in the lack of precise data regarding long-term benefits, the basic principle is that TA could be performed as a routine adjunct to orthotopic heart transplantation by experienced surgeons.

#### *Limitations*

This meta-analysis has some significant limitations. First, it includes three observational retrospective studies: a matched case–control study, two prospective nonrandomized studies, and two RCTs. Second, two of the studies authored by the same team of researchers included the same cohort of patients and reported mostly the same outcomes at different periods, the first after a follow-up of 1 year and the second after a follow-up of 5.7 to 6.7 years. Another group of studies authored by the same authors was conducted respecting identical patient selection criteria and the reported outcomes. To avoid biased results, we have considered the meta-analysis of the data presented in the most recent and representative of them. Third, TR was not uniformly graded in all of the studies. Jeevandaman described four degrees of regurgitation, while the authors used a three-stage classification. Fourthly, there were significant discrepancies regarding the surgical technique. Rubin did not report the technique of heart transplantation at all. The patients included in the study conducted by Brown had undergone biatrial heart transplantation, while the other authors used the bicaval technique. TA was performed by De Vega's technique in all of the studies, except for Brown, who also included the annuloplasties performed using rings.

#### **5. Conclusions**

Our study showed that donor heart tricuspid annuloplasty reduces tricuspid regurgitation incidence in the first year after orthotopic heart transplantation without increasing the surgical complexity. Further large randomized clinical trials are necessary to evaluate the impact of this procedure on long-term insufficiency and outcome benefits. Regarding one-year and long-term mortality, no significant benefit or harm was revealed. Thus, we emphasize the importance of extending the follow-up period on larger cohorts. In conclusion, if performed in high-experienced centers, prophylactic donor tricuspid annuloplasty could be routinely considered during orthotopic heart transplantation as it tends to incline the balance to a more favorable evolution without adding any additional risks.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2227-9 032/9/3/306/s1, Table S1: preferred reporting items for systematic reviews and meta-analysis (PRISMA) checklist, Table S2: Risk of publication bias assessment.

**Author Contributions:** Conceptualization, A.E.B. and A.T.; methodology, A.B..; software, A.E.B.; validation, G.T., A.B. and M.E.; writing—original draft preparation, A.E.B., A.T..; writing—review and editing, A.B.; visualization, M.E..; supervision, G.T.; project administration, A.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


#### *Book Review*

## **Book Review: Cohn, S.L. (Ed.). Decision Making in Perioperative Medicine: Clinical Pearls. (New York: McGraw-Hill), 2021. ISBN: 978-1-260-46810-6**

**Richard H. Parrish II**

Department of Biomedical Sciences, Mercer University School of Medicine, Columbus, GA 31902, USA; parrish\_rh@mercer.edu; Tel.: +1-(706)-321-7218

Cohn's work fills a void in the perioperative care literature by providing a concise, comprehensive, practical, and authoritative guide to the medical management of common periprocedural issues and scenarios. The book is organized logically according to the typical flow of patient management, beginning with an introduction to perioperative patient care, prevention of common complications, treatment of co-existing diseases and special populations (such as the cancer patient and surgery in the older patient), and finishing with the management of common post-operative problems.

The book has a number of strengths, and could be easily converted into a computer app to aid utilization for practice enhancement and teaching at the bedside. It contains many helpful tables and figures (algorithms) that assist the reader to form a complete picture of the topic. The importance of conducting a careful risk assessment to avoid exacerbation of co-morbidities or prevention of various post-operative complications is well highlighted. The 'clinical pearls' found at the end of each chapter provide the reader with high yield and pertinent information on important treatment decision making aspects. The chapters in the section on common post-operative problems are well-written and succinct.

There are also a number of areas where the book could be improved, perhaps in subsequent editions. Because there are 26 co-authors of the various chapters, the scope, depth, and detail of each chapter seems to vary considerably. Some chapters have guideline or clinical trial citations; some important statements were uncited. For example, in the prevention of surgical site infections (SSI) with antimicrobials, the current workhorse antibiotic and dosing for prophylaxis of SSIs, cefazolin, is not mentioned by name; not all cephalosporins are indicated for prophylaxis [1]. This variation is also true of the 'clinical pearls' bullet points, and the source or attribution of the pearl often varies. Some of the pearls are summative statements discussed in-depth in the chapter; others are based on the author's experience or expertise. Moreover, several citations are not the most up to date, such as the management of post-operative nausea and vomiting (PONV) [2]. In this regard, while PONV is a major reason for delayed discharge from the hospital or outpatient surgery center, its assessment and management are scattered among several chapters. It would have been more helpful to the reader if PONV had its own chapter, placed either in the prophylaxis or common problems section. There are several redundancies and omissions regarding medication management. For example, both the chapter on medication management and ischemic heart disease identify cardiac medications that need to be withheld or continued during the perioperative period. Medications to treat myasthenia gravis (pyridostigmine and azathioprine) are omitted in the robust table (chapter 4, page 34), and importance of avoiding general anesthesia and neostigmine reversal in these patients is paramount to safe emergence by using newer agents, perhaps sugammadex (which is not mentioned in the text at all) [3–5]. At times, use of medication names, either brand or generic, is not consistent. There is no chapter on the pediatric patient, which would be another excellent addition to the next edition placed in the special

**Citation:** Parrish II, R.H. Book Review: Cohn, S.L. (Ed.). Decision Making in Perioperative Medicine: Clinical Pearls. (New York: McGraw-Hill), 2021. ISBN: 978-1-260-46810-6. *Healthcare* **2021**, *9*, 687. https://doi.org/10.3390/ healthcare9060687

Academic Editor: Andreas G. Nerlich

Received: 6 May 2021 Accepted: 3 June 2021 Published: 7 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

populations section. Considering the recent major paradigm shift in perioperative care culture, the treatment of enhanced recovery programs is somewhat cursory [6–8].

On balance, while the book has some minor limitations, it is an excellent collection of tips and wisdom proven to be very helpful in the management of the perioperative patient. It is recommended as required reading for any surgery, general medicine, or pharmacy resident or fellow in training or on rotation, as well as a comprehensive reference for experienced practitioners and allied health professionals working in the periprocedural space. It might also serve as the basis of a shared mental model for collaborative practice development among surgeons, anesthesiologists, hospitalists, clinical pharmacists, nurses, dietitians, and other therapists managing the care of the periprocedural patient.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Healthcare* Editorial Office E-mail: healthcare@mdpi.com www.mdpi.com/journal/healthcare

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-1835-0