- Email: [email protected]
CT in critically ill patients with suspected infection: A systematic review and meta-analysis
+
MODEL
Journal of the Formosan Medical Association xxx (xxxx) xxx
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.jfma-online.com
Original Article
Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and metaanalysis Chun-Kai Huang a,b, Jei-Yie Huang b,c, Sheng-Yuan Ruan a, Kuo-Liong Chien b,d,* a Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan b Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan c Department of Nuclear Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan d Division of Cardiology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
Received 10 June 2019; received in revised form 18 September 2019; accepted 20 September 2019
KEYWORDS Infection; Critical care; Fluorodeoxyglucose; Positron emission tomography; Meta-analysis
Background/purpose: Nuclear imaging, including gallium scintigraphy and fluorodeoxyglucose (FDG) positron emission tomography (PET), has been widely used to identify focus of infection in fever of unknown origin. However, little is known about its role in critically ill patients, who are usually with multiple inflammatory foci and unable to tolerate long image acquisition time. This systematic review aimed to evaluate the diagnostic performance of FDG PET for suspected infection in critically ill patients. Methods: PubMed and Embase were searched up to July 24th, 2019 to identify studies evaluating the diagnostic performance of FDG PET for finding infection focus in critically ill patients following the PRISMA guidelines. The bivariate mixed-effects model was used to pool the measure for diagnostic performance. Publication bias was evaluated by Deeks’ method. Results: A total of 4 studies with 87 patients were included. All the four studies evaluated FDG PET. Majority of the patients were either mechanically ventilated (76%) or shocked requiring vasopressors (61%). Test and transportation related adverse events were rare (2%). The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. The AUC for summary ROC curve was 0.83.
* Corresponding author. Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Room 501, No. 17, Xu-Zhou Road, Taipei 100, Taiwan. Fax: þ886 2 2351 1955. E-mail addresses: [email protected] (C.-K. Huang), [email protected] (K.-L. Chien). https://doi.org/10.1016/j.jfma.2019.09.010 0929-6646/Copyright ª 2019, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
2
C.-K. Huang et al. Conclusions: FDG PET was a very sensitive tool with acceptable specificity for detecting the origin of infection in critically ill patients. However, current available studies have limitation in evaluating safety issue. Further research should investigate both benefit and risk of doing this test for this group of vulnerable patients. Copyright ª 2019, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/).
Introduction Infections are the leading cause of death in noncardiac intensive care units (ICUs) and account for 40 percent of all ICU expenditures.1 Sepsis affects millions of people every year and brings remarkable mortality and morbidity impact.2 Early identification of infection focus alters the management of sepsis and improves outcomes.3 A routine work-up for infection, especially sepsis, usually includes microbiological cultures for blood and suspicious sites, biomarkers and image studies such as computed tomography (CT) or sonography. However, previous data showed that more than 50% of severe sepsis was negative for culture and half the culture-negative patients were subsequently proved having infections,4 which reveals the limitation of microbiological cultures in the diagnosis of sepsis and localizing the portal of entry. Gallium scintigraphy had been used to detect inflammatory foci since 1970s5 until the introduction of fluorodeoxyglucose (FDG) positron emission tomography (PET) in the 1990s.6 As gallium accumulates in both malignant tumors and inflammatory tissues, gallium scintigraphy is still used as a component of workup strategies for patients with fever of unknown origin (FUO). FDG PET is a functional imaging modality that can be used for the localization of malignant tumors, as well as infectious and noninfectious inflammatory lesions.7 Several previous studies have assessed the diagnostic usefulness of nuclear imaging tests for localizing the source of infection.7e9 FDG PET/CT had also proven its value in the diagnosis of patients with FUO.10e12 However, critically ill patients could have multiple medical devices and catheters, and complicated inflammatory condition, which may interfere the interpretation of FDG PET. Additionally, their unstable condition often raises safety concern for the transportation to nuclear imaging study. Previous data regarding nuclear imaging in critically ill patients with suspected infection were very limited and subject to small sample size. Data pooling may help to generate more concrete evidence about diagnostic performance and safety issues. Therefore, this systematic review and meta-analysis aimed to pool available evidence to evaluate the diagnostic performance and safety of FDG PET for critically ill patients with suspected infection.
Methods This meta-analysis was performed according to the guidelines of the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA)13 and the Cochrane methodology for diagnostic tool.14 A PRISMA checklist is available in the supplement (see Additional file 1). The detailed review protocol was registered on PROSPERO (registration number: CRD42019120874).
Data sources and search strategy PubMed and Embase were searched (from their inception until July 24th, 2019) with no language restrictions. Prespecified search terms for the target condition were used (including “fever of unknown origin” or “sepsis”) on the target population (including, “critically ill” or “intensive care unit [ICU]”) and the tests of interest (including “gallium scan” and “PET”).14,15 Due to no study evaluated the diagnostic performance of gallium scintigraphy was found, we modified our search terms of tests of interest to “PET” only and repeated the search. The authors perused the reference lists of eligible primary papers, relevant reviews and meta-analyses.
Study selection The article titles and abstracts were reviewed for eligibility. A study was included if it met the following criteria: i) assessed nuclear medicine inflammatory scans for their ability to detect the origin of infection in critically ill patients; ii) the accurate diagnosis of the origin of sepsis was made by culture or further investigation; iii) and absolute numbers of true-positives, false-positives, true-negatives and false-negatives were available, or this data was derivable from the results presented. A study was excluded if it was a case report or case series, or it was a review or metaanalysis, and duplication of patients.
Data extraction and risk assessment of bias and applicability Two researchers (C.K. Huang and J.Y. Huang) independently performed the data extraction. Extracted information included the author, journal, year of publication and country, details of study design; number of patients, mean age, percentage of children, inclusion criteria; imaging modality, brand of imaging device and interpretation method. Disagreement between the 2 researchers was resolved through discussion to a final consensus. We assessed the risk of bias and applicability for each eligible study using items based on the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool (scale,
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management 0e14).16 In brief, the assessment was based on 14 items, including the patient spectrum covered, reference standard, disease progression bias, verification bias, review bias, clinical review bias, incorporation bias, test execution, study withdrawals and indeterminate results. A score of 7e14 indicates a high quality study, and scores below 7 indicate a low quality study. The QUADAS tool is an evidence-based quality assessment tool to be used in systematic reviews of diagnostic accuracy studies.
Data synthesis and statistical analysis For each study, a 2 by 2 contingency table was constructed consisting of true-positive, false-positive, false-negative, and true-negative results; the patients were categorized according to their nuclear imaging test results (positive or negative) and whether the cause of infection was correctly identified by the nuclear imaging test. Categorical variables were expressed as percentages, and continuous variables were expressed as mean values. Pooled measures for diagnostic performance, including sensitivity, specificity, negative predictive value (NPV), diagnostic odds ratios (DORs), summary receiver-operating characteristic (SROC) curves, and area under the curve (AUC) with 95% confidence intervals (CIs) were calculated. True positive scan was categorized as helpful for diagnosis as in Kluge’s study.17 False negative study was defined as failure of FDG PET. Proportion of change of management, helpful for diagnosis and failure of FDG PET were also calculated. The pooled estimates were analyzed using a bivariate mixed-effects regression model to express the pooled diagnostic performance measures across the studies.18e20 The assumption of the bivariate model is that the sensitivities from individual studies (after logit transformation) within a meta-analysis are approximately normally distributed around a mean value, with a certain amount of variability around this mean. The same is true for the specificities of these studies, leading to bivariate normal distribution. The bivariate approach preserves the
3 two-dimensional nature of the original data and incorporates the correlation between these two measures of diagnostic accuracy via random effects. Pooled measures of proportion were also estimated by random-effects models using the score method for calculating confidence intervals due to small sample size.21 Between-study statistical heterogeneity was assessed using I2 and the Cochrane Q test on the basis of the mixed-effects analysis.22 Publication bias was examined using the effective sample size funnel plot and associated regression test of asymmetry as described by Deeks et al.,23 with P < 0.05 for the slope coefficient indicating significant asymmetry. Sepsis in children and adult populations may involve different presentation and clinical consideration. Therefore, we performed subgroup analysis by patient population of children and adult to evaluate the potential effect modification. Sensitivity analysis was demonstrated by DOR using the random effect model after excluding each study in turn.24e26 This is to test the robustness of the summary effect given the circumstance of sparse data in individual study. All analyses were conducted using Stata SE 13.1 (version 14; StataCorp., Texas, USA) and R (version 3.4.2; Foundation for Statistical Computing, Vienna, Austria).
Results Literature flow and study characteristics A total of 33 abstracts were screened and 11 full-text articles were evaluated (Fig. 1). Seven studies were excluded after review of the full article, including 2 case reports, 1 review article, 1 not for infection, 2 without diagnostic performance and 2 studies with overlapping patients.17,27 One study evaluated the diagnostic contribution of gallium scintigraphy but did not reveal information on the sensitivity or specificity due to data complexity in the initial search.28 A total of 4 publications (3 retrospective and 1 prospective, including 87 patients with 89 independent tests) were included for narrative and quantitative
Figure 1 Flow chart of the systematic literature search. A limited search of PubMed and EMBASE yielded 33 articles. After exclusion, a total of 4 studies (87 patients, 89 scans) were included in the final analysis.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
4
C.-K. Huang et al. Table 1
Study characteristics.
Reference Country
Patient number FDG PET Age (SD) timing M/F Mean (SD)
Patients
Modality
Design Study period
Chang 2016 Taiwan
19 5.6 (5.1) 8/11
Critically ill children in PICU, FUO 2 on MV and vasopressor 2 on MV, 2 on vasopressor 13 weaned from MV but requiring oxygen therapy Severe sepsis, no known port of entry for 48 h All on MV
PET/CT (without contrast)
Retro 0.88 2006/07e2014/07 0.67
1. Correlation of CRP and FDG PET 2. Complication not mentioned 3. Pathogen
PET/CT (with contrast)
Pros 0.85 2008/11e2010/10 0.50
1. Transportation and monitoring protocol 2. Complication mentioned 3. Pathogen 1. Transportation and monitoring protocol 2. Monitor glucose every hour 3. No complications 4. FP cause hazardous 1. No complication 2. FP cause hazardous
24.77 (16.31)
Sensitivity Remarks Specificity
Mandry 2014 17 France 51.5 (19.3) 10/7
19.76 (31.07)
Kluge 2012 Germany
18 55.9 (17.8) 14/4
39.94 (28.07)
Severe sepsis or PET/CT septic shock (with 15 on vasopressor contrast) 12 on MV
Retro 1.00 2004/01e2010/12 0.57
Simons 2010 33 Netherland 49.8 (22.4) 24/9
29.06 (27.29)
Suspected infection or inflammatory process in critically ill patients All on MV
PET/CT (without contrast)
Retro 1.00 2005/10e2008/03 0.79
M/F, male/female; PET, positron emission tomography; CT, computed tomography; PICU, pediatric intensive care unit; FUO, fever of unknown origin; MV, mechanically ventilated; Retro, retrospective; Pros, prospective; CRP, C-reactive protein.
synthesis. Table 1 summarizes the characteristics of the included studies.17,29e31 The methodological quality of the 4 studies was assessed using the QUADAS tool. Review of the QUADAS checklist revealed that all of the studies scored between 8 and 10, which indicated good quality (Table 2).16 Most studies were found to have problems with varied time periods between the standard and investigated test, varied reference method for confirming infection origin, dependence between the findings of the investigated and reference tests, lack of reporting for un-interpretable results and a lack of explanation for withdrawals. Verification bias was a major problem in these studies, in which the selection of the reference tests may be greatly influenced by the findings of the index test.
Diagnostic performance, SROC curve, and publication bias Studies of FDG PET produced moderate heterogeneous estimates of sensitivity (I2 Z 38.98%) and homogeneous
estimates of specificity (I2 Z 0%) (Fig. 2). The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. The summary ROC curve revealed an AUC of 0.83 (Fig. 3). The summary NPV was 0.86 (0.47e1.00). Patient 14 in Mandry’s study was categorized as FP and FN in the data synthesis. In this case, endometritis was falsely diagnosed by FDG-PET/CT whereas subsequent Citrobacter. koserii pneumonia was not detected by the FDG PET/CT. This patient was classified as FN in the meta-analysis, as it was considered that endometritis rarely causes severe sepsis. The time span between disease course and FDG PET/ CT was various and long (mean 20e40 days) in the included studies. We analyzed the diagnostic performance divided by binary, tertile and quartile of days of fever (see Additional file 4). Although, there were no association between diagnostic performance and grouping of days of fever, but a trend of fewer false negative results in the longer days of fever group was noted.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management Table 2
5
Quality assessment of each study by QUADAS.
Study
Chang 2016 Mandry 2014 Kluge 2012 Simons 2010
QUADAS tool items
Rating
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Y
N
U
Y Y Y Y
Y Y Y Y
Y Y Y Y
U U U U
N N N N
Y Y Y Y
U Y U U
Y Y Y Y
Y Y Y Y
Y Y Y Y
N U N N
Y Y Y Y
N N N N
U Y U U
8 10 8 8
3 2 3 3
3 2 3 3
Y: yes; N: no; U: unclear.
Figure 2 Forest plot showing the sensitivity and specificity of FDG PET for detecting the origin of sepsis in critically ill patients. The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. I2 Z 38.98% implies moderate heterogeneity in estimating sensitivity and I2 Z 0% implies low heterogeneity in estimating specificity. The size of the grey box provides a measure of the sample size; dots and diamonds represent point estimates; extending lines represent the 95% CI of each estimate. CI, confidence interval; FDG, fluorodeoxyglucose; PET, positron emission tomography.
All 4 studies reported the contribution of FDG PET. However, the definition of contribution was not consistent. Change of management after FDG PET examination was noted in 47% (9/19),29 71% (12/17),30 33% (6/18)17 and 15% (5/33)31 patients. The combined proportion of change of therapy was 0.41 (0.15e0.66). Helpful for final diagnosis was noted in 74% (14/19),29 65% (11/17),30 61% (11/18)17 and 60% (21/35)31 patients. The combined proportion of helpful for diagnosis was 0.65 (0.55e0.74). Failure of FDG PET was noted in 11% (2/19),29 12% (2/17),30 0% (0/18)17 and 0 (0/35)31 patients. The combined proportion of failure of FDG PET was 0.11 (0.01e0.21) by random-effects models. Only Chang and Mandry’s studies reported the causative pathogens and the major pathogens were Candida and Pseudomonas. Reasons for ICU admission were reported in three studies.17,29,31 A total of 4 hepatic failure, 1 heart failure, 1 respiratory failure and 1 multiple organ failure were the ICU admission indication or final outcome in 4 included studies.
The Deeks’ funnel plot and regression test showed a statistically nonsignificant P value (P Z 0.06) for the slope coefficient, indicating symmetric distribution in the data and non-significant publication bias (see Additional file 2).23
Adverse events associated with the test Due to the majority of retrospective nature of the enrolled studies,17,29,31 FDG PET exam failure rate (FDG PET was arranged but was not done due to unstable condition or adverse events) could not be calculated. Mandry et al. and Kluge et al. reported the transportation and monitoring protocol during FDG PET, with at least one physician and one nurse by side during the transportation and all patients were closely monitored (electrocardiography, blood pressure, Saturated O2) during the overall procedure.17,30 All patients fasted for at least 4 h prior to injection with FDG. Parenteral nutrition and all infusions containing glucose or insulin were also discontinued for that period. No episodes
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
6
Figure 3 Summary ROC for detecting septic origin in critically ill patients with FDG PET/CT. The summary ROC curve revealed AUC Z 0.83. Each circle represents individual included study. The diamond represents summary operating point of pooled sensitivity and specificity. The dashed line represents 95% confidence intervals. ROC, receiver operator curve; FDG, fluorodeoxyglucose; PET, positron emission tomography; CT, computed tomography; AUC, area under the curve; SROC, summary receiver-operating characteristic.
of hypoglycemia were noted in any of the studies. Blood glucose concentration was checked before the application of FDG. Administration of FDG was only performed at glucose levels less than 150e200 mg/dL. The blood glucose concentration was also checked every hour before the administration of FDG in the Kluge et al. study.17 In the 87 patients and 89 scans, only two patients had adverse events related to FDG PET, and they were both quickly resolved following treatment. The first was a <90% oxygen desaturation requiring an increase in FiO2, and the second was a drop in systolic arterial pressure <90 mmHg necessitating an increase in the dosage of norepinephrine.30 Kluge et al. and Simons et al. both reported no adverse events, while adverse events and complications were not mentioned in the Chang et al. study.17,29,31 False positive results may lead to unnecessary investigation, which could be hazardous. There were a total of 9 false positive results. Two led to surgical exploration of para-vertebral and tongue lesions,17 while another led to a pericardiocentesis.31 However, the decisions to perform invasive procedures were also based on other examinations, and not the FDG PET alone.
Subgroup analysis Subgroup analysis was differentiated by adults and children with a cutoff age of 18 years. One 17-year old patient who was grouped as an adult in Simon’s study,31 was allocated to
C.-K. Huang et al. the child group in the present study. The mean age (SD) of the adult and child groups were 58.9 (10.2) and 5.9 (7.6) years-old in Simon’s study.31 There were a total of 64 scans for adults (two adults received two FDG PETs) and 25 scans for children. The adult subgroup included 39 true positive, 15 true negative, 8 false positive and 2 false negative scans. The child subgroup had 18 true positive, 4 true negative, 1 false positive and 2 false negative scans. In the adult group, the summary sensitivity and specificity were 0.91 (95% CI, 0.70e0.98) and 0.62 (95% CI, 0.40e0.80), respectively. The SROC curve revealed the AUC was 0.78. In the child group, the summary sensitivity and specificity were 0.86 (95% CI, 0.65e0.96) and 0.70 (95% CI, 0.30e0.93), respectively (see Additional file 3). The SROC curve revealed the AUC was 0.87. FDG PET categorized the detected cause of infection as pneumonia, bone infection, soft tissue infection, vascular related infection, tumor and others. When evaluating for the inflammatory focus in children, an extra category of hematological disease was added. In the adult group the three most commonly detected causes were pneumonia, soft tissue infection and vascular related infection, whereas in the child group bone infection, hematological disease, pneumonia and tumor accounted for most of the causes. It is well known that pneumonia is often difficult to diagnose in patients who are treated by antibiotics and under mechanical ventilation. The positive results from FDG PET played a helpful role to decide when further investigation (such as bronchoalveolar lavage) is needed.
Sensitivity analysis As only a few studies were included within the metaanalysis, sensitivity analysis was performed to find out whether there is an influential study. Sensitivity analysis was performed by excluding each study and pooling the DORs of the remaining 3 studies. Sensitivity analysis after the exclusion of each study showed similar log DORs, which ranged from 2.19 to 3.33 and an overall log DOR of 2.75 (1.32, 4.17). The forest plot of pooled log DOR in the sensitivity study is shown in Fig. 4. The DORs ranged from 8.92 to 28.0.
Discussion We found 4 studies examining the diagnostic accuracy of FDG PET in identifying origin of infection in 87 critically ill patients with suspected infection. The pooled diagnostic performance of FDG PET in this patient population revealed good sensitivity, and acceptable specificity. The area under summary ROC curve was 0.83, indicating good overall diagnostic performance.32 The FDG PET examination subsequently changed patient management in about half of the cases. During the test, majority of the patients were either mechanically ventilated or shocked requiring vasopressors. However, the included studies reported few complications associated with transportation and the test. Generally, these data support that FDG PET is a sensitive and practical imaging modality to identify sepsis origin in critically ill patients.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management
Figure 4 Results of the sensitivity analysis as demonstrated by forest plots of the log diagnostic odds ratio after exclusion of the labeled study. It showed similar log DORs, which ranged from 2.19 to 3.33 and a summary of 2.75 (1.32, 4.17), with no obvious outliers. DOR, diagnostic odds ratio.
Previous meta-analyses had evaluated the diagnostic performance of FDG PET/CT in the assessment of FUO33e35 and revealed a sensitivity 0.98, 0.85, 0.86, a specificity of 0.86, 0.52 and an AUC of 0.95 and 0.88, respectively. Also, FDG PET/CT had the best test performance comparing with other nuclear medicine images.35 There had been some debate on the diagnostic yield of FDG PET/CT for FUO.12,36 However, FDG PET/CT had a high negative predictive value10 and increased in final diagnostic rate.37 Although, there were no previous meta-analyses focusing on critically ill patients for comparison with the present results, our results were similar to previous meta-analyses focusing on FUO in non-critically ill patients, which showed that FDG PET had a high sensitivity, fair specificity and high NPV. The effectiveness of FDG PET in identifying origin of infection in critically ill patients appears to be as good as in usual patients. In our meta-analysis, a trend of fewer false negative results in the longer days of fever group was noted. This trend may also emphasize an important issue of discontinuation or de-escalation of antibiotics treatment in treating critically ill patient with suspected infection. The longer time span between the disease course and FDG PET/CT should be related to prolonged-use of antibiotics. No false negative result was noted when FDG PET/CT was done after 20 days of fever in our meta-analysis. Therefore, if FDG PET/CT showed negative result in these patients, discontinuation or de-escalation of antibiotics treatment could be considered. The majority of studies that evaluated the diagnostic performance of FDG PET for finding the inflammatory focus included adult patients only. In the present study, the child group seemed to show better specificity and AUC. This may be because the majority of the included children were from the study by Chang et al. The patients in Chang’s study were relatively stable, in the sub-acute phase of disease and in the latter period of the diagnostic cascade of FUO, which consequently may have resulted in increased specificity and better AUC in the child group. For septic and critically ill patients, confounding factors such as unstable glucose metabolic status and major organ
7 failure may be an issue. The metabolic status was controlled by fasting and check of blood glucose concentration before FDG PET in all 4 studies. Although, no prior studies precisely evaluated the performance of FDG PET in finding infectious focus in patients with major organ failure (cardiac, respiratory, hepatic, renal), several studies confirm the value FDG PET were worthwhile in impaired major organ function.38e42 Multiple confounding factors in critically ill patients may interfere the interpretation of FDG PET. But, this is real world scenario. This meta-analysis wanted to generate evidence under these circumstances. The present review initially aimed to evaluate evidence on the diagnostic performance and clinical utility of FDG PET and gallium scintigraphy for critically ill patients with suspected infection. However, only one study of gallium scintigraphy was identified and this study did not report the sensitivity and specificity of gallium scintigraphy.28 The reason of rare gallium studies in this patient population may be due to its disadvantages on poor image resolution and longer image acquisition time as compared with FDG PET.43
Strength and limitation This is the first meta-analysis focusing on critically ill patients for evaluating the diagnostic accuracy of FDG PET/ CT. The results for the diagnostic performance of FDG PET/ CT were consistent with the results of previous metaanalysis of FDG PET in FUO patients.35 In addition, we were also concerned about adverse events during the test. We summarized the information about patient condition before the test and adverse event rate during the test. This study has limitations. First, moderate heterogeneity was noted in the data synthesis, which was associated with a small sample size. However, this level of heterogeneity has been the relatively low one compared with previous meta-analyses on this issue.33,34,36,37 Second, the included studies were mainly retrospective study. Information bias is a potential concern for individual study. For example, the risk of adverse events associated with the test may be under-reported and underestimated. Finally, the time span between disease course and FDG PET/CT in the included studies is long. Prolonged-use of antibiotics may affect the diagnostic rate of FDG PET/CT. However, our results did not show more false negative results when longer time span. Moreover, no infectious focus found by FDG PET/CT in patient under antibiotics use implied the effectiveness of antibiotics or well control of the disease.
Conclusions FDG PET/CT is a sensitive diagnostic tool for identifying origin of infection for critically ill patients with suspected infection. The risk of adverse events associated with testing seemed to be low, though the safety issue was not universally reported across all studies. Prospective studies using standardized diagnostic algorithms and safety evaluation are needed to verify the study findings and solid outcomes such as mortality should be used to evaluate the risk and benefit balance for doing FDG PET/CT in this vulnerable patient population.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
8
Declaration of competing interest The authors declare that they have no competing interests. The authors received no financial support for this study. This study was supported, in part, by the Ministry of Science and Technology of Taiwan (106-2314-B-002-049MY3), Taiwan and the National Taiwan University Hospital (108-S4394), Taipei, Taiwan, in English editing and articleprocessing charge.
Acknowledgments We gratefully acknowledge the assistance of Dr. Yu-Kang Tu for the help and advices in statistical analysis. This study was supported in part by the Ministry of Science and Technology of Taiwan (106-2314-B-002 -049 -MY3), Taiwan and the National Taiwan University Hospital (108-S4394), Taipei, Taiwan. No other potential conflict of interest relevant to the study existed.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jfma.2019.09.010.
References 1. Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, et al. International study of the prevalence and outcomes of infection in intensive care units. JAMA 2009;302:2323e9. 2. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et al. Developing a new definition and assessing new clinical criteria for septic shock: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:775e87. 3. Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Rapid increase in hospitalization and mortality rates for severe sepsis in the United States: a trend analysis from 1993 to 2003. Crit Care Med 2007;35:1244e50. 4. Heffner AC, Horton JM, Marchick MR, Jones AE. Etiology of illness in patients with severe sepsis admitted to the hospital from the emergency department. Clin Infect Dis 2010;50: 814e20. 5. Forgacs P. The gallium scan and inflammatory lesions. JAMA Intern Med 1982;142:231. 6. Hicks RJ, Hofman MS. Is there still a role for SPECT-CT in oncology in the PET-CT era? Nat Rev Clin Oncol 2012;9:712e20. 7. Knockaert DC, Vanderschueren S, Blockmans D. Fever of unknown origin in adults: 40 years on. J Intern Med 2003;253: 263e75. 8. Mourad O, Palda V, Detsky AS. A comprehensive evidencebased approach to fever of unknown origin. Arch Intern Med 2003;163:545e51. 9. Varghese GM, Trowbridge P, Doherty T. Investigating and managing pyrexia of unknown origin in adults. BMJ 2010;341: 999e1000. 10. Revest M, Patrat-Delon S, Devillers A, Tattevin P, Michelet C. Contribution of 18fluoro-deoxyglucose PET/CT for the diagnosis of infectious diseases. Med Mal Infect 2014;44:251e60. 11. Kouijzer IJE, Mulders-Manders CM, Bleeker-Rovers CP, Oyen WJG. Fever of unknown origin: the value of FDG-PET/CT. Semin Nucl Med 2018;48:100e7.
C.-K. Huang et al. 12. Kouijzer IJE, van der Meer JWM, Oyen WJG, Bleeker-Rovers CP. Diagnostic yield of FDG-PET/CT in fever of unknown origin: a systematic review, meta-analysis, and Delphi exercise. Clin Radiol 2018;73:588e9. 13. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009;339:b2700. 14. Deeks JJBP, Gatsonis C. Cochrane handbook for systematic reviews of diagnostic test accuracy. The Cochrane Collaboration; 2013. Version 1.0.0. 15. de Vet HCW EA, Riphagen II, Aertgeerts B, Pewsner D. Chapter 7: Searching for studies. In: Cochrane handbook for systematic reviews of diagnostic test accuracy. The Cochrane Collaboration; 2008. Version 04. 16. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25. 17. Kluge S, Braune S, Nierhaus A, Wichmann D, Derlin T, Mester J, et al. Diagnostic value of positron emission tomography combined with computed tomography for evaluating patients with septic shock of unknown origin. J Crit Care 2012;27:316. e1e7. 18. Chu H, Cole SR. Bivariate meta-analysis of sensitivity and specificity with sparse data: a generalized linear mixed model approach. J Clin Epidemiol 2006;59:1331e3. 19. Hamza TH, van Houwelingen HC, Stijnen T. The binomial distribution of meta-analysis was preferred to model within-study variability. J Clin Epidemiol 2008;61:41e51. 20. Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005;58:982e90. 21. Nyaga VN, Arbyn M, Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Arch Public Health 2014;72:39. 22. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ 1994;309:1351e5. 23. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882e93. 24. Doebler P, Holling H. Meta-analysis of diagnostic accuracy with mada. 2015. Reterieved at: https://cran.r.project.org/web/ packages/mada/vignettes/mada.pdf. 25. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The diagnostic odds ratio: a single indicator of test performance. J Clin Epidemiol 2003;56:1129e35. 26. DerSimonian R, Laird N. Meta-analysis in clinical trials. Contr Clin Trials 1986;7:177e88. 27. Kampe KK, Rotermund R, Tienken M, Thomalla G, Regier M, Klutmann S, et al. Diagnostic value of positron emission tomography combined with computed tomography for evaluating critically ill neurological patients. Front Neurol 2017;8:33. 28. Chen WC, Tsai KD, Chen CH, Lin MS, Chen CM, Shih CM, et al. Role of gallium-67 scintigraphy in the evaluation of occult sepsis in the medical ICU. Intern Emerg Med 2012;7: 53e8. 29. Chang L, Cheng MF, Jou ST, Ko CL, Huang JY, Tzen KY, et al. Search of unknown fever focus using PET in critically ill children with complicated underlying diseases. Pediatr Crit Care Med 2016;17:e58e65. 30. Mandry D, Tatopoulos A, Chevalier-Mathias E, Lemarie J, Bollaert PE, Roch V, et al. (1)(8)F-fluorodeoxyglucose positron emission tomography combined with whole-body computed tomographic angiography in critically ill patients with
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management
31.
32. 33.
34.
35.
36.
37.
suspected severe sepsis with no definite diagnosis. Eur J Nucl Med Mol Imaging 2014;41:1924e30. Simons KS, Pickkers P, Bleeker-Rovers CP, Oyen WJ, van der Hoeven JG. F-18-fluorodeoxyglucose positron emission tomography combined with CT in critically ill patients with suspected infection. Intensive Care Med 2010;36:504e11. Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8:283e98. Dong MJ, Zhao K, Liu ZF, Wang GL, Yang SY, Zhou GJ. A metaanalysis of the value of fluorodeoxyglucose-PET/PET-CT in the evaluation of fever of unknown origin. Eur J Radiol 2011;80: 834e44. Hao R, Yuan L, Kan Y, Li C, Yang J. Diagnostic performance of 18F-FDG PET/CT in patients with fever of unknown origin: a meta-analysis. Nucl Med Commun 2013;34:682e8. Takeuchi M, Dahabreh IJ, Nihashi T, Iwata M, Varghese GM, Terasawa T. Nuclear imaging for classic fever of unknown origin: meta-analysis. J Nucl Med 2016;57:1913e9. Bharucha T, Rutherford A, Skeoch S, Alavi A, Brown M, Galloway J. Diagnostic yield of FDG-PET/CT in fever of unknown origin: a systematic review, meta-analysis, and Delphi exercise. Clin Radiol 2017;72:764e71. Besson FL, Chaumet-Riffaud P, Playe M, Noel N, Lambotte O, Goujard C, et al. Contribution of (18)F-FDG PET in the
9
38.
39.
40. 41. 42.
43.
diagnostic assessment of fever of unknown origin (FUO): a stratification-based meta-analysis. Eur J Nucl Med Mol Imaging 2016;43:1887e95. Fan H-B, Wang A-J, Yang D-L, Xiao J, Ai Y, Huang L, et al. Use of (18)F-FDG PET/CT to locate primary malignancies in patients with hepatic cirrhosis and malignant ascites. Chin J Canc Res 2013;25:500e4. Toriihara A, Kitazume Y, Nishida H, Kubota K, Nakadate M, Tateishi U. Comparison of FDG-PET/CT images between chronic renal failure patients on hemodialysis and controls. Am J Nucl Med Mol Imaging 2015;5:204e11. Kode V, Karsch H, Osman MM, Muzaffar R. Impact of renal failure on F18-FDG PET/CT scans. Front Oncol 2017;7:155. Quail MA, Sinusas AJ. PET-CMR in heart failure - synergistic or redundant imaging? Heart Fail Rev 2017;22:477e89. Vankadari K, Mittal BR, Kumar R, Singh H, Bhattacharya A, Dhiman RK. Detection of hepatic encephalopathy on 18F-FDG PET/CT brain images in a patient with decompensated liver cirrhosis. Clin Nucl Med 2018;43:e486e7. Palestro CJ, Love C. Nuclear medicine imaging in fever of unknown origin: the new Paradigm. Curr Pharmaceut Des 2018; 24:814e20.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
MODEL
Journal of the Formosan Medical Association xxx (xxxx) xxx
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.jfma-online.com
Original Article
Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and metaanalysis Chun-Kai Huang a,b, Jei-Yie Huang b,c, Sheng-Yuan Ruan a, Kuo-Liong Chien b,d,* a Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan b Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan c Department of Nuclear Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan d Division of Cardiology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
Received 10 June 2019; received in revised form 18 September 2019; accepted 20 September 2019
KEYWORDS Infection; Critical care; Fluorodeoxyglucose; Positron emission tomography; Meta-analysis
Background/purpose: Nuclear imaging, including gallium scintigraphy and fluorodeoxyglucose (FDG) positron emission tomography (PET), has been widely used to identify focus of infection in fever of unknown origin. However, little is known about its role in critically ill patients, who are usually with multiple inflammatory foci and unable to tolerate long image acquisition time. This systematic review aimed to evaluate the diagnostic performance of FDG PET for suspected infection in critically ill patients. Methods: PubMed and Embase were searched up to July 24th, 2019 to identify studies evaluating the diagnostic performance of FDG PET for finding infection focus in critically ill patients following the PRISMA guidelines. The bivariate mixed-effects model was used to pool the measure for diagnostic performance. Publication bias was evaluated by Deeks’ method. Results: A total of 4 studies with 87 patients were included. All the four studies evaluated FDG PET. Majority of the patients were either mechanically ventilated (76%) or shocked requiring vasopressors (61%). Test and transportation related adverse events were rare (2%). The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. The AUC for summary ROC curve was 0.83.
* Corresponding author. Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Room 501, No. 17, Xu-Zhou Road, Taipei 100, Taiwan. Fax: þ886 2 2351 1955. E-mail addresses: [email protected] (C.-K. Huang), [email protected] (K.-L. Chien). https://doi.org/10.1016/j.jfma.2019.09.010 0929-6646/Copyright ª 2019, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
2
C.-K. Huang et al. Conclusions: FDG PET was a very sensitive tool with acceptable specificity for detecting the origin of infection in critically ill patients. However, current available studies have limitation in evaluating safety issue. Further research should investigate both benefit and risk of doing this test for this group of vulnerable patients. Copyright ª 2019, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/).
Introduction Infections are the leading cause of death in noncardiac intensive care units (ICUs) and account for 40 percent of all ICU expenditures.1 Sepsis affects millions of people every year and brings remarkable mortality and morbidity impact.2 Early identification of infection focus alters the management of sepsis and improves outcomes.3 A routine work-up for infection, especially sepsis, usually includes microbiological cultures for blood and suspicious sites, biomarkers and image studies such as computed tomography (CT) or sonography. However, previous data showed that more than 50% of severe sepsis was negative for culture and half the culture-negative patients were subsequently proved having infections,4 which reveals the limitation of microbiological cultures in the diagnosis of sepsis and localizing the portal of entry. Gallium scintigraphy had been used to detect inflammatory foci since 1970s5 until the introduction of fluorodeoxyglucose (FDG) positron emission tomography (PET) in the 1990s.6 As gallium accumulates in both malignant tumors and inflammatory tissues, gallium scintigraphy is still used as a component of workup strategies for patients with fever of unknown origin (FUO). FDG PET is a functional imaging modality that can be used for the localization of malignant tumors, as well as infectious and noninfectious inflammatory lesions.7 Several previous studies have assessed the diagnostic usefulness of nuclear imaging tests for localizing the source of infection.7e9 FDG PET/CT had also proven its value in the diagnosis of patients with FUO.10e12 However, critically ill patients could have multiple medical devices and catheters, and complicated inflammatory condition, which may interfere the interpretation of FDG PET. Additionally, their unstable condition often raises safety concern for the transportation to nuclear imaging study. Previous data regarding nuclear imaging in critically ill patients with suspected infection were very limited and subject to small sample size. Data pooling may help to generate more concrete evidence about diagnostic performance and safety issues. Therefore, this systematic review and meta-analysis aimed to pool available evidence to evaluate the diagnostic performance and safety of FDG PET for critically ill patients with suspected infection.
Methods This meta-analysis was performed according to the guidelines of the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA)13 and the Cochrane methodology for diagnostic tool.14 A PRISMA checklist is available in the supplement (see Additional file 1). The detailed review protocol was registered on PROSPERO (registration number: CRD42019120874).
Data sources and search strategy PubMed and Embase were searched (from their inception until July 24th, 2019) with no language restrictions. Prespecified search terms for the target condition were used (including “fever of unknown origin” or “sepsis”) on the target population (including, “critically ill” or “intensive care unit [ICU]”) and the tests of interest (including “gallium scan” and “PET”).14,15 Due to no study evaluated the diagnostic performance of gallium scintigraphy was found, we modified our search terms of tests of interest to “PET” only and repeated the search. The authors perused the reference lists of eligible primary papers, relevant reviews and meta-analyses.
Study selection The article titles and abstracts were reviewed for eligibility. A study was included if it met the following criteria: i) assessed nuclear medicine inflammatory scans for their ability to detect the origin of infection in critically ill patients; ii) the accurate diagnosis of the origin of sepsis was made by culture or further investigation; iii) and absolute numbers of true-positives, false-positives, true-negatives and false-negatives were available, or this data was derivable from the results presented. A study was excluded if it was a case report or case series, or it was a review or metaanalysis, and duplication of patients.
Data extraction and risk assessment of bias and applicability Two researchers (C.K. Huang and J.Y. Huang) independently performed the data extraction. Extracted information included the author, journal, year of publication and country, details of study design; number of patients, mean age, percentage of children, inclusion criteria; imaging modality, brand of imaging device and interpretation method. Disagreement between the 2 researchers was resolved through discussion to a final consensus. We assessed the risk of bias and applicability for each eligible study using items based on the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool (scale,
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management 0e14).16 In brief, the assessment was based on 14 items, including the patient spectrum covered, reference standard, disease progression bias, verification bias, review bias, clinical review bias, incorporation bias, test execution, study withdrawals and indeterminate results. A score of 7e14 indicates a high quality study, and scores below 7 indicate a low quality study. The QUADAS tool is an evidence-based quality assessment tool to be used in systematic reviews of diagnostic accuracy studies.
Data synthesis and statistical analysis For each study, a 2 by 2 contingency table was constructed consisting of true-positive, false-positive, false-negative, and true-negative results; the patients were categorized according to their nuclear imaging test results (positive or negative) and whether the cause of infection was correctly identified by the nuclear imaging test. Categorical variables were expressed as percentages, and continuous variables were expressed as mean values. Pooled measures for diagnostic performance, including sensitivity, specificity, negative predictive value (NPV), diagnostic odds ratios (DORs), summary receiver-operating characteristic (SROC) curves, and area under the curve (AUC) with 95% confidence intervals (CIs) were calculated. True positive scan was categorized as helpful for diagnosis as in Kluge’s study.17 False negative study was defined as failure of FDG PET. Proportion of change of management, helpful for diagnosis and failure of FDG PET were also calculated. The pooled estimates were analyzed using a bivariate mixed-effects regression model to express the pooled diagnostic performance measures across the studies.18e20 The assumption of the bivariate model is that the sensitivities from individual studies (after logit transformation) within a meta-analysis are approximately normally distributed around a mean value, with a certain amount of variability around this mean. The same is true for the specificities of these studies, leading to bivariate normal distribution. The bivariate approach preserves the
3 two-dimensional nature of the original data and incorporates the correlation between these two measures of diagnostic accuracy via random effects. Pooled measures of proportion were also estimated by random-effects models using the score method for calculating confidence intervals due to small sample size.21 Between-study statistical heterogeneity was assessed using I2 and the Cochrane Q test on the basis of the mixed-effects analysis.22 Publication bias was examined using the effective sample size funnel plot and associated regression test of asymmetry as described by Deeks et al.,23 with P < 0.05 for the slope coefficient indicating significant asymmetry. Sepsis in children and adult populations may involve different presentation and clinical consideration. Therefore, we performed subgroup analysis by patient population of children and adult to evaluate the potential effect modification. Sensitivity analysis was demonstrated by DOR using the random effect model after excluding each study in turn.24e26 This is to test the robustness of the summary effect given the circumstance of sparse data in individual study. All analyses were conducted using Stata SE 13.1 (version 14; StataCorp., Texas, USA) and R (version 3.4.2; Foundation for Statistical Computing, Vienna, Austria).
Results Literature flow and study characteristics A total of 33 abstracts were screened and 11 full-text articles were evaluated (Fig. 1). Seven studies were excluded after review of the full article, including 2 case reports, 1 review article, 1 not for infection, 2 without diagnostic performance and 2 studies with overlapping patients.17,27 One study evaluated the diagnostic contribution of gallium scintigraphy but did not reveal information on the sensitivity or specificity due to data complexity in the initial search.28 A total of 4 publications (3 retrospective and 1 prospective, including 87 patients with 89 independent tests) were included for narrative and quantitative
Figure 1 Flow chart of the systematic literature search. A limited search of PubMed and EMBASE yielded 33 articles. After exclusion, a total of 4 studies (87 patients, 89 scans) were included in the final analysis.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
4
C.-K. Huang et al. Table 1
Study characteristics.
Reference Country
Patient number FDG PET Age (SD) timing M/F Mean (SD)
Patients
Modality
Design Study period
Chang 2016 Taiwan
19 5.6 (5.1) 8/11
Critically ill children in PICU, FUO 2 on MV and vasopressor 2 on MV, 2 on vasopressor 13 weaned from MV but requiring oxygen therapy Severe sepsis, no known port of entry for 48 h All on MV
PET/CT (without contrast)
Retro 0.88 2006/07e2014/07 0.67
1. Correlation of CRP and FDG PET 2. Complication not mentioned 3. Pathogen
PET/CT (with contrast)
Pros 0.85 2008/11e2010/10 0.50
1. Transportation and monitoring protocol 2. Complication mentioned 3. Pathogen 1. Transportation and monitoring protocol 2. Monitor glucose every hour 3. No complications 4. FP cause hazardous 1. No complication 2. FP cause hazardous
24.77 (16.31)
Sensitivity Remarks Specificity
Mandry 2014 17 France 51.5 (19.3) 10/7
19.76 (31.07)
Kluge 2012 Germany
18 55.9 (17.8) 14/4
39.94 (28.07)
Severe sepsis or PET/CT septic shock (with 15 on vasopressor contrast) 12 on MV
Retro 1.00 2004/01e2010/12 0.57
Simons 2010 33 Netherland 49.8 (22.4) 24/9
29.06 (27.29)
Suspected infection or inflammatory process in critically ill patients All on MV
PET/CT (without contrast)
Retro 1.00 2005/10e2008/03 0.79
M/F, male/female; PET, positron emission tomography; CT, computed tomography; PICU, pediatric intensive care unit; FUO, fever of unknown origin; MV, mechanically ventilated; Retro, retrospective; Pros, prospective; CRP, C-reactive protein.
synthesis. Table 1 summarizes the characteristics of the included studies.17,29e31 The methodological quality of the 4 studies was assessed using the QUADAS tool. Review of the QUADAS checklist revealed that all of the studies scored between 8 and 10, which indicated good quality (Table 2).16 Most studies were found to have problems with varied time periods between the standard and investigated test, varied reference method for confirming infection origin, dependence between the findings of the investigated and reference tests, lack of reporting for un-interpretable results and a lack of explanation for withdrawals. Verification bias was a major problem in these studies, in which the selection of the reference tests may be greatly influenced by the findings of the index test.
Diagnostic performance, SROC curve, and publication bias Studies of FDG PET produced moderate heterogeneous estimates of sensitivity (I2 Z 38.98%) and homogeneous
estimates of specificity (I2 Z 0%) (Fig. 2). The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. The summary ROC curve revealed an AUC of 0.83 (Fig. 3). The summary NPV was 0.86 (0.47e1.00). Patient 14 in Mandry’s study was categorized as FP and FN in the data synthesis. In this case, endometritis was falsely diagnosed by FDG-PET/CT whereas subsequent Citrobacter. koserii pneumonia was not detected by the FDG PET/CT. This patient was classified as FN in the meta-analysis, as it was considered that endometritis rarely causes severe sepsis. The time span between disease course and FDG PET/ CT was various and long (mean 20e40 days) in the included studies. We analyzed the diagnostic performance divided by binary, tertile and quartile of days of fever (see Additional file 4). Although, there were no association between diagnostic performance and grouping of days of fever, but a trend of fewer false negative results in the longer days of fever group was noted.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management Table 2
5
Quality assessment of each study by QUADAS.
Study
Chang 2016 Mandry 2014 Kluge 2012 Simons 2010
QUADAS tool items
Rating
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Y
N
U
Y Y Y Y
Y Y Y Y
Y Y Y Y
U U U U
N N N N
Y Y Y Y
U Y U U
Y Y Y Y
Y Y Y Y
Y Y Y Y
N U N N
Y Y Y Y
N N N N
U Y U U
8 10 8 8
3 2 3 3
3 2 3 3
Y: yes; N: no; U: unclear.
Figure 2 Forest plot showing the sensitivity and specificity of FDG PET for detecting the origin of sepsis in critically ill patients. The summary sensitivity and specificity were 0.94 (95% CI, 0.79e0.99) and 0.66 (95% CI, 0.45e0.83), respectively. I2 Z 38.98% implies moderate heterogeneity in estimating sensitivity and I2 Z 0% implies low heterogeneity in estimating specificity. The size of the grey box provides a measure of the sample size; dots and diamonds represent point estimates; extending lines represent the 95% CI of each estimate. CI, confidence interval; FDG, fluorodeoxyglucose; PET, positron emission tomography.
All 4 studies reported the contribution of FDG PET. However, the definition of contribution was not consistent. Change of management after FDG PET examination was noted in 47% (9/19),29 71% (12/17),30 33% (6/18)17 and 15% (5/33)31 patients. The combined proportion of change of therapy was 0.41 (0.15e0.66). Helpful for final diagnosis was noted in 74% (14/19),29 65% (11/17),30 61% (11/18)17 and 60% (21/35)31 patients. The combined proportion of helpful for diagnosis was 0.65 (0.55e0.74). Failure of FDG PET was noted in 11% (2/19),29 12% (2/17),30 0% (0/18)17 and 0 (0/35)31 patients. The combined proportion of failure of FDG PET was 0.11 (0.01e0.21) by random-effects models. Only Chang and Mandry’s studies reported the causative pathogens and the major pathogens were Candida and Pseudomonas. Reasons for ICU admission were reported in three studies.17,29,31 A total of 4 hepatic failure, 1 heart failure, 1 respiratory failure and 1 multiple organ failure were the ICU admission indication or final outcome in 4 included studies.
The Deeks’ funnel plot and regression test showed a statistically nonsignificant P value (P Z 0.06) for the slope coefficient, indicating symmetric distribution in the data and non-significant publication bias (see Additional file 2).23
Adverse events associated with the test Due to the majority of retrospective nature of the enrolled studies,17,29,31 FDG PET exam failure rate (FDG PET was arranged but was not done due to unstable condition or adverse events) could not be calculated. Mandry et al. and Kluge et al. reported the transportation and monitoring protocol during FDG PET, with at least one physician and one nurse by side during the transportation and all patients were closely monitored (electrocardiography, blood pressure, Saturated O2) during the overall procedure.17,30 All patients fasted for at least 4 h prior to injection with FDG. Parenteral nutrition and all infusions containing glucose or insulin were also discontinued for that period. No episodes
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
6
Figure 3 Summary ROC for detecting septic origin in critically ill patients with FDG PET/CT. The summary ROC curve revealed AUC Z 0.83. Each circle represents individual included study. The diamond represents summary operating point of pooled sensitivity and specificity. The dashed line represents 95% confidence intervals. ROC, receiver operator curve; FDG, fluorodeoxyglucose; PET, positron emission tomography; CT, computed tomography; AUC, area under the curve; SROC, summary receiver-operating characteristic.
of hypoglycemia were noted in any of the studies. Blood glucose concentration was checked before the application of FDG. Administration of FDG was only performed at glucose levels less than 150e200 mg/dL. The blood glucose concentration was also checked every hour before the administration of FDG in the Kluge et al. study.17 In the 87 patients and 89 scans, only two patients had adverse events related to FDG PET, and they were both quickly resolved following treatment. The first was a <90% oxygen desaturation requiring an increase in FiO2, and the second was a drop in systolic arterial pressure <90 mmHg necessitating an increase in the dosage of norepinephrine.30 Kluge et al. and Simons et al. both reported no adverse events, while adverse events and complications were not mentioned in the Chang et al. study.17,29,31 False positive results may lead to unnecessary investigation, which could be hazardous. There were a total of 9 false positive results. Two led to surgical exploration of para-vertebral and tongue lesions,17 while another led to a pericardiocentesis.31 However, the decisions to perform invasive procedures were also based on other examinations, and not the FDG PET alone.
Subgroup analysis Subgroup analysis was differentiated by adults and children with a cutoff age of 18 years. One 17-year old patient who was grouped as an adult in Simon’s study,31 was allocated to
C.-K. Huang et al. the child group in the present study. The mean age (SD) of the adult and child groups were 58.9 (10.2) and 5.9 (7.6) years-old in Simon’s study.31 There were a total of 64 scans for adults (two adults received two FDG PETs) and 25 scans for children. The adult subgroup included 39 true positive, 15 true negative, 8 false positive and 2 false negative scans. The child subgroup had 18 true positive, 4 true negative, 1 false positive and 2 false negative scans. In the adult group, the summary sensitivity and specificity were 0.91 (95% CI, 0.70e0.98) and 0.62 (95% CI, 0.40e0.80), respectively. The SROC curve revealed the AUC was 0.78. In the child group, the summary sensitivity and specificity were 0.86 (95% CI, 0.65e0.96) and 0.70 (95% CI, 0.30e0.93), respectively (see Additional file 3). The SROC curve revealed the AUC was 0.87. FDG PET categorized the detected cause of infection as pneumonia, bone infection, soft tissue infection, vascular related infection, tumor and others. When evaluating for the inflammatory focus in children, an extra category of hematological disease was added. In the adult group the three most commonly detected causes were pneumonia, soft tissue infection and vascular related infection, whereas in the child group bone infection, hematological disease, pneumonia and tumor accounted for most of the causes. It is well known that pneumonia is often difficult to diagnose in patients who are treated by antibiotics and under mechanical ventilation. The positive results from FDG PET played a helpful role to decide when further investigation (such as bronchoalveolar lavage) is needed.
Sensitivity analysis As only a few studies were included within the metaanalysis, sensitivity analysis was performed to find out whether there is an influential study. Sensitivity analysis was performed by excluding each study and pooling the DORs of the remaining 3 studies. Sensitivity analysis after the exclusion of each study showed similar log DORs, which ranged from 2.19 to 3.33 and an overall log DOR of 2.75 (1.32, 4.17). The forest plot of pooled log DOR in the sensitivity study is shown in Fig. 4. The DORs ranged from 8.92 to 28.0.
Discussion We found 4 studies examining the diagnostic accuracy of FDG PET in identifying origin of infection in 87 critically ill patients with suspected infection. The pooled diagnostic performance of FDG PET in this patient population revealed good sensitivity, and acceptable specificity. The area under summary ROC curve was 0.83, indicating good overall diagnostic performance.32 The FDG PET examination subsequently changed patient management in about half of the cases. During the test, majority of the patients were either mechanically ventilated or shocked requiring vasopressors. However, the included studies reported few complications associated with transportation and the test. Generally, these data support that FDG PET is a sensitive and practical imaging modality to identify sepsis origin in critically ill patients.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management
Figure 4 Results of the sensitivity analysis as demonstrated by forest plots of the log diagnostic odds ratio after exclusion of the labeled study. It showed similar log DORs, which ranged from 2.19 to 3.33 and a summary of 2.75 (1.32, 4.17), with no obvious outliers. DOR, diagnostic odds ratio.
Previous meta-analyses had evaluated the diagnostic performance of FDG PET/CT in the assessment of FUO33e35 and revealed a sensitivity 0.98, 0.85, 0.86, a specificity of 0.86, 0.52 and an AUC of 0.95 and 0.88, respectively. Also, FDG PET/CT had the best test performance comparing with other nuclear medicine images.35 There had been some debate on the diagnostic yield of FDG PET/CT for FUO.12,36 However, FDG PET/CT had a high negative predictive value10 and increased in final diagnostic rate.37 Although, there were no previous meta-analyses focusing on critically ill patients for comparison with the present results, our results were similar to previous meta-analyses focusing on FUO in non-critically ill patients, which showed that FDG PET had a high sensitivity, fair specificity and high NPV. The effectiveness of FDG PET in identifying origin of infection in critically ill patients appears to be as good as in usual patients. In our meta-analysis, a trend of fewer false negative results in the longer days of fever group was noted. This trend may also emphasize an important issue of discontinuation or de-escalation of antibiotics treatment in treating critically ill patient with suspected infection. The longer time span between the disease course and FDG PET/CT should be related to prolonged-use of antibiotics. No false negative result was noted when FDG PET/CT was done after 20 days of fever in our meta-analysis. Therefore, if FDG PET/CT showed negative result in these patients, discontinuation or de-escalation of antibiotics treatment could be considered. The majority of studies that evaluated the diagnostic performance of FDG PET for finding the inflammatory focus included adult patients only. In the present study, the child group seemed to show better specificity and AUC. This may be because the majority of the included children were from the study by Chang et al. The patients in Chang’s study were relatively stable, in the sub-acute phase of disease and in the latter period of the diagnostic cascade of FUO, which consequently may have resulted in increased specificity and better AUC in the child group. For septic and critically ill patients, confounding factors such as unstable glucose metabolic status and major organ
7 failure may be an issue. The metabolic status was controlled by fasting and check of blood glucose concentration before FDG PET in all 4 studies. Although, no prior studies precisely evaluated the performance of FDG PET in finding infectious focus in patients with major organ failure (cardiac, respiratory, hepatic, renal), several studies confirm the value FDG PET were worthwhile in impaired major organ function.38e42 Multiple confounding factors in critically ill patients may interfere the interpretation of FDG PET. But, this is real world scenario. This meta-analysis wanted to generate evidence under these circumstances. The present review initially aimed to evaluate evidence on the diagnostic performance and clinical utility of FDG PET and gallium scintigraphy for critically ill patients with suspected infection. However, only one study of gallium scintigraphy was identified and this study did not report the sensitivity and specificity of gallium scintigraphy.28 The reason of rare gallium studies in this patient population may be due to its disadvantages on poor image resolution and longer image acquisition time as compared with FDG PET.43
Strength and limitation This is the first meta-analysis focusing on critically ill patients for evaluating the diagnostic accuracy of FDG PET/ CT. The results for the diagnostic performance of FDG PET/ CT were consistent with the results of previous metaanalysis of FDG PET in FUO patients.35 In addition, we were also concerned about adverse events during the test. We summarized the information about patient condition before the test and adverse event rate during the test. This study has limitations. First, moderate heterogeneity was noted in the data synthesis, which was associated with a small sample size. However, this level of heterogeneity has been the relatively low one compared with previous meta-analyses on this issue.33,34,36,37 Second, the included studies were mainly retrospective study. Information bias is a potential concern for individual study. For example, the risk of adverse events associated with the test may be under-reported and underestimated. Finally, the time span between disease course and FDG PET/CT in the included studies is long. Prolonged-use of antibiotics may affect the diagnostic rate of FDG PET/CT. However, our results did not show more false negative results when longer time span. Moreover, no infectious focus found by FDG PET/CT in patient under antibiotics use implied the effectiveness of antibiotics or well control of the disease.
Conclusions FDG PET/CT is a sensitive diagnostic tool for identifying origin of infection for critically ill patients with suspected infection. The risk of adverse events associated with testing seemed to be low, though the safety issue was not universally reported across all studies. Prospective studies using standardized diagnostic algorithms and safety evaluation are needed to verify the study findings and solid outcomes such as mortality should be used to evaluate the risk and benefit balance for doing FDG PET/CT in this vulnerable patient population.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
8
Declaration of competing interest The authors declare that they have no competing interests. The authors received no financial support for this study. This study was supported, in part, by the Ministry of Science and Technology of Taiwan (106-2314-B-002-049MY3), Taiwan and the National Taiwan University Hospital (108-S4394), Taipei, Taiwan, in English editing and articleprocessing charge.
Acknowledgments We gratefully acknowledge the assistance of Dr. Yu-Kang Tu for the help and advices in statistical analysis. This study was supported in part by the Ministry of Science and Technology of Taiwan (106-2314-B-002 -049 -MY3), Taiwan and the National Taiwan University Hospital (108-S4394), Taipei, Taiwan. No other potential conflict of interest relevant to the study existed.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jfma.2019.09.010.
References 1. Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, et al. International study of the prevalence and outcomes of infection in intensive care units. JAMA 2009;302:2323e9. 2. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et al. Developing a new definition and assessing new clinical criteria for septic shock: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:775e87. 3. Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Rapid increase in hospitalization and mortality rates for severe sepsis in the United States: a trend analysis from 1993 to 2003. Crit Care Med 2007;35:1244e50. 4. Heffner AC, Horton JM, Marchick MR, Jones AE. Etiology of illness in patients with severe sepsis admitted to the hospital from the emergency department. Clin Infect Dis 2010;50: 814e20. 5. Forgacs P. The gallium scan and inflammatory lesions. JAMA Intern Med 1982;142:231. 6. Hicks RJ, Hofman MS. Is there still a role for SPECT-CT in oncology in the PET-CT era? Nat Rev Clin Oncol 2012;9:712e20. 7. Knockaert DC, Vanderschueren S, Blockmans D. Fever of unknown origin in adults: 40 years on. J Intern Med 2003;253: 263e75. 8. Mourad O, Palda V, Detsky AS. A comprehensive evidencebased approach to fever of unknown origin. Arch Intern Med 2003;163:545e51. 9. Varghese GM, Trowbridge P, Doherty T. Investigating and managing pyrexia of unknown origin in adults. BMJ 2010;341: 999e1000. 10. Revest M, Patrat-Delon S, Devillers A, Tattevin P, Michelet C. Contribution of 18fluoro-deoxyglucose PET/CT for the diagnosis of infectious diseases. Med Mal Infect 2014;44:251e60. 11. Kouijzer IJE, Mulders-Manders CM, Bleeker-Rovers CP, Oyen WJG. Fever of unknown origin: the value of FDG-PET/CT. Semin Nucl Med 2018;48:100e7.
C.-K. Huang et al. 12. Kouijzer IJE, van der Meer JWM, Oyen WJG, Bleeker-Rovers CP. Diagnostic yield of FDG-PET/CT in fever of unknown origin: a systematic review, meta-analysis, and Delphi exercise. Clin Radiol 2018;73:588e9. 13. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009;339:b2700. 14. Deeks JJBP, Gatsonis C. Cochrane handbook for systematic reviews of diagnostic test accuracy. The Cochrane Collaboration; 2013. Version 1.0.0. 15. de Vet HCW EA, Riphagen II, Aertgeerts B, Pewsner D. Chapter 7: Searching for studies. In: Cochrane handbook for systematic reviews of diagnostic test accuracy. The Cochrane Collaboration; 2008. Version 04. 16. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25. 17. Kluge S, Braune S, Nierhaus A, Wichmann D, Derlin T, Mester J, et al. Diagnostic value of positron emission tomography combined with computed tomography for evaluating patients with septic shock of unknown origin. J Crit Care 2012;27:316. e1e7. 18. Chu H, Cole SR. Bivariate meta-analysis of sensitivity and specificity with sparse data: a generalized linear mixed model approach. J Clin Epidemiol 2006;59:1331e3. 19. Hamza TH, van Houwelingen HC, Stijnen T. The binomial distribution of meta-analysis was preferred to model within-study variability. J Clin Epidemiol 2008;61:41e51. 20. Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005;58:982e90. 21. Nyaga VN, Arbyn M, Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Arch Public Health 2014;72:39. 22. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ 1994;309:1351e5. 23. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882e93. 24. Doebler P, Holling H. Meta-analysis of diagnostic accuracy with mada. 2015. Reterieved at: https://cran.r.project.org/web/ packages/mada/vignettes/mada.pdf. 25. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The diagnostic odds ratio: a single indicator of test performance. J Clin Epidemiol 2003;56:1129e35. 26. DerSimonian R, Laird N. Meta-analysis in clinical trials. Contr Clin Trials 1986;7:177e88. 27. Kampe KK, Rotermund R, Tienken M, Thomalla G, Regier M, Klutmann S, et al. Diagnostic value of positron emission tomography combined with computed tomography for evaluating critically ill neurological patients. Front Neurol 2017;8:33. 28. Chen WC, Tsai KD, Chen CH, Lin MS, Chen CM, Shih CM, et al. Role of gallium-67 scintigraphy in the evaluation of occult sepsis in the medical ICU. Intern Emerg Med 2012;7: 53e8. 29. Chang L, Cheng MF, Jou ST, Ko CL, Huang JY, Tzen KY, et al. Search of unknown fever focus using PET in critically ill children with complicated underlying diseases. Pediatr Crit Care Med 2016;17:e58e65. 30. Mandry D, Tatopoulos A, Chevalier-Mathias E, Lemarie J, Bollaert PE, Roch V, et al. (1)(8)F-fluorodeoxyglucose positron emission tomography combined with whole-body computed tomographic angiography in critically ill patients with
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010
+
MODEL
FDG PET in infection management
31.
32. 33.
34.
35.
36.
37.
suspected severe sepsis with no definite diagnosis. Eur J Nucl Med Mol Imaging 2014;41:1924e30. Simons KS, Pickkers P, Bleeker-Rovers CP, Oyen WJ, van der Hoeven JG. F-18-fluorodeoxyglucose positron emission tomography combined with CT in critically ill patients with suspected infection. Intensive Care Med 2010;36:504e11. Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8:283e98. Dong MJ, Zhao K, Liu ZF, Wang GL, Yang SY, Zhou GJ. A metaanalysis of the value of fluorodeoxyglucose-PET/PET-CT in the evaluation of fever of unknown origin. Eur J Radiol 2011;80: 834e44. Hao R, Yuan L, Kan Y, Li C, Yang J. Diagnostic performance of 18F-FDG PET/CT in patients with fever of unknown origin: a meta-analysis. Nucl Med Commun 2013;34:682e8. Takeuchi M, Dahabreh IJ, Nihashi T, Iwata M, Varghese GM, Terasawa T. Nuclear imaging for classic fever of unknown origin: meta-analysis. J Nucl Med 2016;57:1913e9. Bharucha T, Rutherford A, Skeoch S, Alavi A, Brown M, Galloway J. Diagnostic yield of FDG-PET/CT in fever of unknown origin: a systematic review, meta-analysis, and Delphi exercise. Clin Radiol 2017;72:764e71. Besson FL, Chaumet-Riffaud P, Playe M, Noel N, Lambotte O, Goujard C, et al. Contribution of (18)F-FDG PET in the
9
38.
39.
40. 41. 42.
43.
diagnostic assessment of fever of unknown origin (FUO): a stratification-based meta-analysis. Eur J Nucl Med Mol Imaging 2016;43:1887e95. Fan H-B, Wang A-J, Yang D-L, Xiao J, Ai Y, Huang L, et al. Use of (18)F-FDG PET/CT to locate primary malignancies in patients with hepatic cirrhosis and malignant ascites. Chin J Canc Res 2013;25:500e4. Toriihara A, Kitazume Y, Nishida H, Kubota K, Nakadate M, Tateishi U. Comparison of FDG-PET/CT images between chronic renal failure patients on hemodialysis and controls. Am J Nucl Med Mol Imaging 2015;5:204e11. Kode V, Karsch H, Osman MM, Muzaffar R. Impact of renal failure on F18-FDG PET/CT scans. Front Oncol 2017;7:155. Quail MA, Sinusas AJ. PET-CMR in heart failure - synergistic or redundant imaging? Heart Fail Rev 2017;22:477e89. Vankadari K, Mittal BR, Kumar R, Singh H, Bhattacharya A, Dhiman RK. Detection of hepatic encephalopathy on 18F-FDG PET/CT brain images in a patient with decompensated liver cirrhosis. Clin Nucl Med 2018;43:e486e7. Palestro CJ, Love C. Nuclear medicine imaging in fever of unknown origin: the new Paradigm. Curr Pharmaceut Des 2018; 24:814e20.
Please cite this article as: Huang C-K et al., Diagnostic performance of FDG PET/CT in critically ill patients with suspected infection: A systematic review and meta-analysis, Journal of the Formosan Medical Association, https://doi.org/10.1016/j.jfma.2019.09.010