Skip to main content
PMC home page
Journal of Clinical Microbiology logoLink to Journal of Clinical Microbiology
. 2017 Aug 23;55(9):2679–2685. doi: 10.1128/JCM.00267-17

Rapid Differentiation of Haemophilus influenzae and Haemophilus haemolyticus by Use of Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry with ClinProTools Mass Spectrum Analysis

Jonathan H K Chen a, Vincent C C Cheng a,b, Chun-Pong Wong a, Sally C Y Wong a, Wing-Cheong Yam a, Kwok-Yung Yuen a,c,✉
Editor: Yi-Wei Tangd
PMCID: PMC5648705  PMID: 28637909

ABSTRACT

Haemophilus influenzae is associated with severe invasive disease, while Haemophilus haemolyticus is considered part of the commensal flora in the human respiratory tract. Although the addition of a custom mass spectrum library into the matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) system could improve identification of these two species, the establishment of such a custom database is technically complicated and requires a large amount of resources, which most clinical laboratories cannot afford. In this study, we developed a mass spectrum analysis model with 7 mass peak biomarkers for the identification of H. influenzae and H. haemolyticus using the ClinProTools software. We evaluated the diagnostic performance of this model using 408 H. influenzae and H. haemolyticus isolates from clinical respiratory specimens from 363 hospitalized patients and compared the identification results with those obtained with the Bruker IVD MALDI Biotyper. The IVD MALDI Biotyper identified only 86.9% of H. influenzae (311/358) and 98.0% of H. haemolyticus (49/50) clinical isolates to the species level. In comparison, the ClinProTools mass spectrum model could identify 100% of H. influenzae (358/358) and H. haemolyticus (50/50) clinical strains to the species level and significantly improved the species identification rate (McNemar's test, P < 0.0001). In conclusion, the use of ClinProTools demonstrated an alternative way for users lacking special expertise in mass spectrometry to handle closely related bacterial species when the proprietary spectrum library failed. This approach should be useful for the differentiation of other closely related bacterial species.

KEYWORDS: MALDI-TOF, Haemophilus haemolyticus, Haemophilus influenzae, ClinProTools

INTRODUCTION

Haemophilus influenzae and Haemophilus haemolyticus are the most prevalent Haemophilus species colonizing the human pharyngeal cavity (1). H. influenzae causes various respiratory tract diseases, such as sinusitis, otitis media, and pneumonia, as well as infective exacerbations of chronic obstructive pulmonary disease and bronchiectasis (2). In contrast, H. haemolyticus colonizes healthy subjects at moderate prevalence and rarely causes infections in the respiratory tract (3–5).

Laboratory differentiation of the two species based on phenotypic tests can be difficult. H. haemolyticus and H. influenzae demonstrate similar colony morphologies, and both require hemin (X factor) and NAD (V factor) for growth. Hemolytic activity on rabbit or horse blood agar was previously used as an important factor for the differentiation of these two species, but this method had been shown to be unreliable, and loss of hemolytic activity of H. haemolyticus after subculture can happen (1, 6, 7). Molecular diagnosis through 16S rRNA gene sequencing or PCR with specific genes, such as the Haemophilus protein D gene (hpd) or fuculose kinase gene (fucK), have been shown to be reliable alternatives. The prevalence of either or both of the hpd and fucK genes in H. influenzae can be up to 92% (8–11). However, molecular testing is expensive and technically demanding.

Recently, proteomic profiling by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), an alternative to biochemical and molecular identification methods, has been evaluated in species differentiation for a variety of microorganisms, with promising results (12, 13). However, some closely related bacterial species could not be well differentiated by the MALDI Biotyper (Bruker Daltonics, Bremen, Germany) alone, and additional mass spectrum analyses would be required for complete resolution (14–17). Although previous studies demonstrated that MALDI-TOF MS could be used to differentiate H. influenzae from H. haemolyticus, H. influenzae strains from China were subsequently reported to have protein mass spectra different from those of strains isolated from other areas (18). Additional local Haemophilus spectrum entries to the manufacturer's reference library would be necessary to optimize the accuracy of species identification (19).

In addition to the use of MALDI Biotyper alone for bacterial identification, mass spectrum analysis through the use of the Bruker ClinProTools software (Bruker Daltonics, Bremen, Germany) has been introduced for the detection of critical biomarker mass peaks, and to develop classification models for the identification of certain phylogenetically related bacterial species in a rapid and user-friendly manner (14).

In this study, we aimed to establish a reliable rapid diagnostic method for the differentiation of H. influenzae and H. haemolyticus from clinical isolates using MALDI-TOF MS with ClinProTools mass spectrum model analysis.

RESULTS

Routine biochemical and molecular identification.

All 408 Haemophilus species isolates were oxidase-positive Gram-negative rods demonstrating X and V factor dependence on Columbia horse blood agar (CHBA). Hemolysis was observed on CHBA within 48 h of incubation in 46 out of the 408 isolates. Of the 408 strains, 314 were hpd positive by PCR. Among the 94 hpd-negative isolates, 44 were indicated fucK positive by PCR. These 358 (314 hpd positive and 44 hpd and fucK positive) isolates were considered H. influenzae. The remaining 50 hpd- and fucK-negative isolates were subjected to 16S rRNA gene sequencing, and all of them were identified as H. haemolyticus through NCBI BLAST. These 50 H. haemolyticus isolates included the 46 strains showing hemolysis on CHBA.

IVD MALDI Biotyper identification alone.

The stand-alone IVD MALDI Biotyper yielded concordant results for 360/408 (88.2%) Haemophilus strains, including 311/358 (86.9%) H. influenzae and 49/50 (98.0%) H. haemolyticus strains, compared to the routine identification method. An indeterminate result was obtained for 48/408 (11.8%) isolates, where both H. influenzae and H. haemolyticus with scores of >2.0 were included in the IVD MALDI Biotyper top 10 identification matches (Table 1). According to the standard operating procedure of the Bruker IVD MALDI Biotyper system, these 48 isolates could only be confirmed to the genus level as Haemophilus species, and further tests would be necessary for species confirmation.

TABLE 1.

Characteristics of the 48 Haemophilus isolates with undetermined species identification using stand-alone MALDI Biotypera

Isolate no. Result by:
Biochemical tests
 Molecular tests
 MALDI Biotyper
 ClinProTools model
X and V factor dependent Hemolysis hpd fucK Confirmed ID Highest score ID (Bruker score) Other species within the top 10 IDs (Bruker score)
4 Y N N P H. influenzae H. influenzae (2.316) H. haemolyticus (2.224) H. influenzae
6 Y N P ND H. influenzae H. influenzae (2.235) H. haemolyticus (2.186) H. influenzae
12 Y N P ND H. influenzae H. influenzae (2.212) H. haemolyticus (2.14) H. influenzae
22 Y N P ND H. influenzae H. influenzae (2.257) H. haemolyticus (2.199) H. influenzae
23 Y N N P H. influenzae H. influenzae (2.315) H. haemolyticus (2.24) H. influenzae
24 Y N P ND H. influenzae H. haemolyticus (2.207) H. influenzae (2.203) H. influenzae
26 Y N P ND H. influenzae H. influenzae (2.358) H. haemolyticus (2.283) H. influenzae
27 Y N P ND H. influenzae H. influenzae (2.289) H. haemolyticus (2.275) H. influenzae
39 Y N N P H. influenzae H. influenzae (2.227) H. haemolyticus (2.185) H. influenzae
42 Y N P ND H. influenzae H. influenzae (2.233) H. haemolyticus (2.107) H. influenzae
44 Y N P ND H. influenzae H. influenzae (2.382) H. haemolyticus (2.244) H. influenzae
56 Y N P ND H. influenzae H. influenzae (2.392) H. haemolyticus (2.234) H. influenzae
57 Y N P ND H. influenzae H. influenzae (2.277) H. haemolyticus (2.194) H. influenzae
64 Y N P ND H. influenzae H. influenzae (2.314) H. haemolyticus (2.254) H. influenzae
70 Y N P ND H. influenzae H. influenzae (2.357) H. haemolyticus (2.273) H. influenzae
78 Y Y N N H. haemolyticusb H. haemolyticus (2.294) H. influenzae (2.212) H. haemolyticus
91 Y N P ND H. influenzae H. influenzae (2.388) H. haemolyticus (2.244) H. influenzae
96 Y N N P H. influenzae H. influenzae (2.234) H. haemolyticus (2.189) H. influenzae
99 Y N P ND H. influenzae H. influenzae (2.243) H. haemolyticus (2.174) H. influenzae
106 Y N P ND H. influenzae H. influenzae (2.328) H. haemolyticus (2.182) H. influenzae
117 Y N P ND H. influenzae H. influenzae (2.418) H. haemolyticus (2.241) H. influenzae
123 Y N P ND H. influenzae H. influenzae (2.384) H. haemolyticus (2.192) H. influenzae
124 Y N P ND H. influenzae H. influenzae (2.274) H. haemolyticus (2.144) H. influenzae
126 Y N P ND H. influenzae H. influenzae (2.233) H. haemolyticus (2.098) H. influenzae
129 Y N P ND H. influenzae H. influenzae (2.272) H. haemolyticus (2.184) H. influenzae
147 Y N P ND H. influenzae H. influenzae (2.352) H. haemolyticus (2.234) H. influenzae
148 Y N P ND H. influenzae H. influenzae (2.38) H. haemolyticus (2.205) H. influenzae
157 Y N P ND H. influenzae H. influenzae (2.298) H. haemolyticus (2.199) H. influenzae
159 Y N P ND H. influenzae H. influenzae (2.353) H. haemolyticus (2.26) H. influenzae
169 Y N P ND H. influenzae H. influenzae (2.367) H. haemolyticus (2.234) H. influenzae
170 Y N P ND H. influenzae H. influenzae (2.313) H. haemolyticus (2.239) H. influenzae
171 Y N P ND H. influenzae H. influenzae (2.274) H. haemolyticus (2.213) H. influenzae
172 Y N N P H. influenzae H. influenzae (2.268) H. haemolyticus (2.238) H. influenzae
246 Y N P ND H. influenzae H. influenzae (2.256) H. haemolyticus (2.187) H. influenzae
255 Y N P ND H. influenzae H. influenzae (2.229) H. haemolyticus (2.205) H. influenzae
267 Y N N P H. influenzae H. influenzae (2.205) H. haemolyticus (2.085) H. influenzae
277 Y N P ND H. influenzae H. influenzae (2.259) H. haemolyticus (2.205) H. influenzae
278 Y N P ND H. influenzae H. influenzae (2.418) H. haemolyticus (2.339) H. influenzae
284 Y N P ND H. influenzae H. influenzae (2.282) H. haemolyticus (2.236) H. influenzae
324 Y N P ND H. influenzae H. haemolyticus (2.292) H. influenzae (2.289) H. influenzae
336 Y N P ND H. influenzae H. influenzae (2.37) H. haemolyticus (2.22) H. influenzae
340 Y N P ND H. influenzae H. influenzae (2.386) H. haemolyticus (2.288) H. influenzae
348 Y N P ND H. influenzae H. influenzae (2.333) H. haemolyticus (2.272) H. influenzae
359 Y N P ND H. influenzae H. influenzae (2.269) H. haemolyticus (2.207) H. influenzae
360 Y N P ND H. influenzae H. influenzae (2.379) H. haemolyticus (2.186) H. influenzae
364 Y N P ND H. influenzae H. influenzae (2.346) H. haemolyticus (2.191) H. influenzae
369 Y N P ND H. influenzae H. haemolyticus (2.157) H. influenzae (2.136) H. influenzae
375 Y N P ND H. influenzae H. influenzae (2.376) H. haemolyticus (2.236) H. influenzae
Open in a new tab
a

ID, identity; Y, yes; P, positive; N, negative; ND, not determined.

b

The identification was further confirmed by 16S rRNA gene sequencing.

ClinProTools model analysis.

Due to the incomplete identification of the IVD MALDI Biotyper alone, the use of mass peak analysis with ClinProTools was explored for the complete solution to differentiate H. influenzae from H. haemolyticus. In ClinProTools, a classification model based on 480 spectra of H. influenzae or H. haemolyticus reference strains was developed using the support vector machine (SVM) algorithm. The classification model included 7 mass peaks in the range of 4,459.73 Da to 9,421.92 Da which demonstrated 100% recognition capability (RC) and cross validation (CV). These 7 mass peaks had area under receiver operating characteristic curve (AUC) values of 0.959 to 1.000 (Table 2). This model could successfully classify all 20 H. influenzae and 20 H. haemolyticus strains in the validation panel. The other 10 H. parainfluenzae and 10 H. parahaemolyticus strains were classified as indeterminate, which was defined as neither H. influenzae nor H. haemolyticus. For clinical isolates, this model correctly classified all 408 Haemophilus isolates, including 358 H. influenzae and 50 H. haemolyticus isolates, to the species level. The ClinProTools mass spectrum model analysis significantly improved the successful identification rate, in comparison to the stand-alone MALDI Biotyper, from 88.2% to 100% (McNemar's test, P < 0.0001).

TABLE 2.

ClinProTools mass peak statistics for the 7 peaks included in the SVM modela

Index Mass (m/z) DAve PTTA PWKW PAD Ave1 Ave2 SD1 SD2 CV1 CV2 AUC
1 4,459.73 8.99 <0.000001 <0.000001 <0.000001 0.72 9.71 0.17 5.9 23.59 60.81 0.994
2 4,590.36 2.53 <0.000001 0 <0.000001 3.42 0.89 1.76 0.36 51.49 40.93 0.988
3 4,711.49 16.19 <0.000001 0 <0.000001 22.58 6.4 8.15 5.22 36.07 81.6 0.959
4 4,746.97 11.37 <0.000001 <0.000001 <0.000001 5.15 16.52 3.36 3.44 65.3 20.81 0.987
5 8,918.2 17.31 <0.000001 <0.000001 <0.000001 1.29 18.59 0.2 9.76 15.32 52.51 1.000
6 9,179.4 2.76 <0.000001 0 <0.000001 3.59 0.83 1.46 0.43 40.53 51.16 0.989
7 9,421.92 27.74 <0.000001 0 <0.000001 39.56 11.82 11.58 6.4 29.28 54.09 0.991
Open in a new tab
a

DAve, difference between the maximal and the minimal average peak area/intensity of all classes; PTTA, P value of t test; PWKW, P value of Wilcoxon test; PAD, P value of Anderson-Darling test; Avg1 and Avg2, peak area/intensity averages of class 1 (H. influenzae) and class 2 (H. haemolyticus), respectively; SD1 and SD2, standard deviations of the peak area/intensity average of class 1 (H. influenzae) and class 2 (H. haemolyticus), respectively; CV1 and CV2, coefficients of variation (percent) of class 1 (H. influenzae) and class 2 (H. haemolyticus), respectively; AUC, area under receiver operating characteristic curve.

DISCUSSION

Discrimination of closely related Haemophilus species is of diagnostic importance because of their differences in pathogenicity. Morphological differentiation and biochemical tests are still the standard method used in many laboratories. In our study, we included 358 H. influenzae and 50 H. haemolyticus clinical isolates. Among the 50 H. haemolyticus isolates confirmed by 16S rRNA gene sequencing, 4 isolates had no demonstrable hemolytic activity. This highlighted the presence of nonhemolytic H. haemolyticus in our locality. Our data are consistent with the findings from previous studies in Denmark, indicating that the lack of hemolysis should not be used as the critical factor to differentiate H. haemolyticus from H. influenzae (6, 7). Although the molecular method by specific gene PCR was proven to be superior to biochemical identification (9–11, 20), the long processing time and high running cost made this difficult to be used in routine clinical diagnostic services.

In this study, we proved the concept of using a mass peak analysis software to help laboratory users lacking expertise in mass spectrometry to differentiate closely related bacterial species. We compared the performance of IVD MALDI Biotyper and ClinProTools peak analysis with that of stand-alone MALDI-TOF MS in the differentiation of H. influenzae and H. haemolyticus. Despite the fact that the most updated MALDI Biotyper reference library includes 27 H. influenzae and 21 H. haemolyticus mass spectrum profiles, our study demonstrated that a significant number of clinical isolates in our region, 13.1% (47/358) of H. influenzae isolates and 2.0% (1/50) of H. haemolyticus isolates, could not be identified to the species level by MALDI Biotyper alone. Our results are similar to those reported in a previous publication from China (18). This indicates that the expansion of reference library provided by the manufacturer cannot resolve the identification problem posed by H. influenzae and H. haemolyticus isolates.

We have developed an SVM mass spectrum analysis model using the ClinProTools software that delivers a practical solution to this current problem. With respect to the classification model, it has to include both reference and clinical strains with confirmed identities. The mass spectra should be prepared by using the ethanol-formic acid extraction method, in order to minimize the noise interference in the mass spectra. The 7 most critical species-specific mass peaks were manually selected into the model mainly based on the AUC values of the peaks. The mass peaks with m/z at 4,590.36 Da, 4,711.49 Da, 9,179.40 Da, and 9,421.92 Da were found to be H. influenzae specific, while the mass peaks with m/z at 4,459.73 Da, 4,746.97 Da, and 8,918.20 Da were H. haemolyticus specific (Table 3). Although both the ClinProTools and the IVD MALDI Biotyper analyze the same set of spectral data, the ClinProTools method focusing on the 7 species-specific peaks was demonstrated to have better accuracy than the IVD MALDI Biotyper. By using the ClinProTools SVM model, the rate of successful H. influenzae and H. haemolyticus identification increased from 88.2% to 100%. This should be closely related to the different analyzing methods used by the two methods.

TABLE 3.

Summary of MALDI-TOF MS results of the 408 Haemophilus clinical isolates

Reference identification (biochemical and molecular method) Species-level identification, no. (%)
MALDI Biotyper alonea ClinProTools
H. influenzae (n = 358) 311 (86.9) 358 (100)
H. haemolyticus (n = 50) 49 (98.0) 50 (100)
Open in a new tab
a

Another 48 isolates contained both H. influenzae and H. haemolyticus in their top 10 identification matches in Bruker MALDI Biotyper.

On the other hand, if only the top-scoring Biotyper identification result was picked and the other identifications with scores of ≥2.0 were neglected, 407/408 (99.8%) isolates could also give correct species-level identifications in the study cohort. However, this “picking rule” has not been statistically evaluated, and thus, there may be a potential risk of misidentification.

The working procedure for the ClinProTools analysis was simple and user-friendly. The hands-on time from importing the spectra file into the ClinProTools software, data analysis, and report generation was approximately 5 min. The recommended workflow would be an initial identification using the IVD MALDI Biotyper system. Isolates with identification of H. influenzae or H. haemolyticus should be further analyzed by the ClinProTools model analysis for species confirmation.

This study has a few limitations. Since this mass spectrum analysis model is specifically designed for H. influenzae and H. haemolyticus differentiation, it is not suitable for analyzing species other than H. influenzae and H. haemolyticus. Although the study demonstrated the feasibility of location-specific improvements to MALDI-TOF MS identification, the ClinProTools method may demonstrate different diagnostic performance for Haemophilus strains from different geographical regions. Further evaluation will be required. In conclusion, our study demonstrated the use of the ClinProTools mass spectrum model analysis, which is a reliable, simple, and cost-effective clinical diagnostic method for the differentiation of H. influenzae and H. haemolyticus. This approach would also be useful for the differentiation of other closely related bacterial species.

MATERIALS AND METHODS

Specimen collection.

A total of 408 Haemophilus isolates, including 358 H. influenzae and 50 H. haemolyticus isolates from 363 patients hospitalized between March 2016 and December 2016, were selected for the study. These isolates originated from various clinical specimen types, including 366 sputum specimens, 12 bronchoalveolar lavage specimens, 13 tracheal or bronchial aspirates, and 17 miscellaneous swabs collected from the ear, eye, or anterior nares. The specimens were cultured on Columbia horse blood agar (CHBA) and chocolate agar, incubated at 37°C with 5% CO2 for 48 h. Isolated Gram-negative rod colonies with small, smooth, and translucent colony morphologies suggestive of Haemophilus species were subjected to species identification through biochemical tests, molecular tests, MALDI-TOF MS, and ClinProTools mass spectrum model analysis (21).

Reference biochemical and molecular identification method.

Isolated colonies were identified by colony morphology, Gram staining, and basic biochemical tests, including oxidase test, sugar fermentation, and hemin (X factor) and NAD (V factor) dependence. Beta-hemolytic activity was also evaluated by culturing strains on 5% CHBA. Haemophilus strains demonstrating beta-hemolysis within 48 h of incubation are considered beta-hemolytic, but hemolytic effect after 48 h is regarded as nonhemolytic.

For identity confirmation, two H. influenzae-specific genes, hpd and fucK, were used in a sequential manner. First, the presence of hpd was examined in all 408 isolates by PCR with primers hpd751F (5′-AAAGACCCAAAGGGTTATTGGGTA-3′) and hpd965R (5′-TTACGCACGGTGTAAGGATGCAC-3′). For hpd-negative isolates, the presence of fucK was further examined by fucK-up (5′-ACCACTTTCGGCGTGGATGG-3′) and fucK-dn (5′-AAGATTTCCCAGGTGCCAGA-3′) (http://www.mlst.net). Isolates that were hpd or fucK positive and X and V factor dependent were considered to be H. influenzae. On the other hand, hpd and fucK PCR-negative, X and V factor-dependent isolates would be considered H. haemolyticus (9). For isolates with ambiguous PCR results, partial 16S rRNA gene sequencing was performed (8, 22).

Species identification by MALDI Biotyper IVD 2.3.

All 408 clinical strains were prepared for MALDI-TOF MS using the manufacturer's recommended direct transfer protocol. A single isolated colony of each strain was inoculated directly onto the MSP96 target plate spot. Each spot was overlaid with α-cyano-4-hydroxy-cinnamic acid matrix (Sigma-Aldrich, St. Louis, MO). The target plate was analyzed by the Bruker microflex LT system (Bruker Daltonics, Bremen, Germany). The protein profile of each spot with m/z values of 3,000 to 15,000 was analyzed by the IVD MALDI Biotyper version 2.3 with the spectrum library, V.5.0.0.0 (5,989 spectra). The top 10 identification matches were generated along with confidence scores, ranging from 0.0 to 3.0. According to the manufacturer's CE-IVD certified recommendations, a score of >2.0 indicates promising species-level identification. If two or more different species were shown within the top 10 matches, the species of the isolate would be considered indeterminate.

Species differentiation by ClinProTools.

Bruker ClinProTools version 3.0 software was used to develop a classification model for rapid differentiation of H. influenzae and H. haemolyticus, with modification from a previous protocol (14). Ten well-characterized H. influenzae (2 ATCC standard strains and 8 well-characterized clinical strains) and 10 H. haemolyticus clinical reference strains were selected for the construction of a classification model. For each strain, 24 high-quality spectra were prepared using the ethanol formic acid extraction method and were captured using flexControl version 3.4 software (Bruker Daltonics, Bremen, Germany). A mass spectrum model consisting of 480 mass spectra from the 20 H. influenzae or H. haemolyticus reference strains was created in the ClinProTools. Spectrum pretreatment and peak calculation operations were performed using the preset configuration. A classification model was generated using the support vector machine (SVM) algorithm. The number of peaks selected in the model was limited to 10, and only peaks with area under receiver operating characteristic curve (AUC) values of >0.9 were considered (23). The recognition capability (RC) and cross validation (CV) percentages of the SVM model were generated to demonstrate the reliability and accuracy of the model. RC and CV percentages were indicators of the model's performance and were useful predictors of the model's ability to classify test isolates. The model was then validated blindly by 3 laboratory technicians with archived clinical strains: 20 H. influenzae isolates, 20 H. haemolyticus isolates, and 20 other Haemophilus isolates (10 H. parainfluenzae and 10 H. parahaemolyticus isolates). The model showing 100% accuracy was selected for further testing. For the 408 clinical strains, the corresponding spectra created by the direct transfer method during the IVD MALDI Biotyper identification were imported into ClinProTools and were further analyzed with the SVM classification model for H. influenzae and H. haemolyticus differentiation.

Statistical analysis.

The identification accuracies of IVD MALDI Biotyper alone and IVD MALDI Biotyper plus ClinProTools analysis were compared using chi-square test and McNemar's test in MedCalc software version 14.12.0 (MedCalc Software, Ostend, Belgium).

ACKNOWLEDGMENTS

We thank Siddharth Sridhar for his valuable comments on the manuscript and the clinical laboratory technicians of Department of Microbiology, Queen Mary Hospital, for providing the clinical isolates.

We declare no conflict of interest.

REFERENCES

  • 1.McCrea KW, Xie J, LaCross N, Patel M, Mukundan D, Murphy TF, Marrs CF, Gilsdorf JR. 2008. Relationships of nontypeable Haemophilus influenzae strains to hemolytic and nonhemolytic Haemophilus haemolyticus strains. J Clin Microbiol 46:406–416. doi: 10.1128/JCM.01832-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Rosell A, Monso E, Soler N, Torres F, Angrill J, Riise G, Zalacain R, Morera J, Torres A. 2005. Microbiologic determinants of exacerbation in chronic obstructive pulmonary disease. Arch Intern Med 165:891–897. doi: 10.1001/archinte.165.8.891. [DOI] [PubMed] [Google Scholar]
  • 3.Anderson R, Wang X, Briere EC, Katz LS, Cohn AC, Clark TA, Messonnier NE, Mayer LW. 2012. Haemophilus haemolyticus isolates causing clinical disease. J Clin Microbiol 50:2462–2465. doi: 10.1128/JCM.06575-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Murphy TF. 2003. Respiratory infections caused by non-typeable Haemophilus influenzae. Curr Opin Infect Dis 16:129–134. doi: 10.1097/00001432-200304000-00009. [DOI] [PubMed] [Google Scholar]
  • 5.Nørskov-Lauritsen N. 2014. Classification, identification, and clinical significance of Haemophilus and Aggregatibacter species with host specificity for humans. Clin Microbiol Rev 27:214–240. doi: 10.1128/CMR.00103-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Fenger MG, Ridderberg W, Olesen HV, Norskov-Lauritsen N. 2012. Low occurrence of 'non-haemolytic Haemophilus haemolyticus' misidentified as Haemophilus influenzae in cystic fibrosis respiratory specimens, and frequent recurrence of persistent H. influenzae clones despite antimicrobial treatment. Int J Med Microbiol 302:315–319. doi: 10.1016/j.ijmm.2012.10.001. [DOI] [PubMed] [Google Scholar]
  • 7.Murphy TF, Brauer AL, Sethi S, Kilian M, Cai X, Lesse AJ. 2007. Haemophilus haemolyticus: a human respiratory tract commensal to be distinguished from Haemophilus influenzae. J Infect Dis 195:81–89. doi: 10.1086/509824. [DOI] [PubMed] [Google Scholar]
  • 8.Janda JM, Abbott SL. 2007. 16S rRNA gene sequencing for bacterial identification in the diagnostic laboratory: pluses, perils, and pitfalls. J Clin Microbiol 45:2761–2764. doi: 10.1128/JCM.01228-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Theodore MJ, Anderson RD, Wang X, Katz LS, Vuong JT, Bell ME, Juni BA, Lowther SA, Lynfield R, MacNeil JR, Mayer LW. 2012. Evaluation of new biomarker genes for differentiating Haemophilus influenzae from Haemophilus haemolyticus. J Clin Microbiol 50:1422–1424. doi: 10.1128/JCM.06702-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Nørskov-Lauritsen N. 2009. Detection of cryptic genospecies misidentified as Haemophilus influenzae in routine clinical samples by assessment of marker genes fucK, hap, and sodC. J Clin Microbiol 47:2590–2592. doi: 10.1128/JCM.00013-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Pickering J, Binks MJ, Beissbarth J, Hare KM, Kirkham LA, Smith-Vaughan H. 2014. A PCR-high-resolution melt assay for rapid differentiation of nontypeable Haemophilus influenzae and Haemophilus haemolyticus. J Clin Microbiol 52:663–667. doi: 10.1128/JCM.02191-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Luo Y, Siu GK, Yeung AS, Chen JH, Ho PL, Leung KW, Tsang JL, Cheng VC, Guo L, Yang J, Ye L, Yam WC. 2015. Performance of the VITEK MS matrix-assisted laser desorption ionization-time of flight mass spectrometry system for rapid bacterial identification in two diagnostic centres in China. J Med Microbiol 64:18–24. doi: 10.1099/jmm.0.080317-0. [DOI] [PubMed] [Google Scholar]
  • 13.Chen JH, Yam WC, Ngan AH, Fung AM, Woo WL, Yan MK, Choi GK, Ho PL, Cheng VC, Yuen KY. 2013. Advantages of using matrix-assisted laser desorption ionization–time of flight mass spectrometry as a rapid diagnostic tool for identification of yeasts and mycobacteria in the clinical microbiological laboratory. J Clin Microbiol 51:3981–3987. doi: 10.1128/JCM.01437-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chen JH, She KK, Wong OY, Teng JL, Yam WC, Lau SK, Woo PC, Cheng VC, Yuen KY. 2015. Use of MALDI Biotyper plus ClinProTools mass spectra analysis for correct identification of Streptococcus pneumoniae and Streptococcus mitis/oralis. J Clin Pathol 68:652–656. doi: 10.1136/jclinpath-2014-202818. [DOI] [PubMed] [Google Scholar]
  • 15.Teng SH, Chen CM, Lee MR, Lee TF, Chien KY, Teng LJ, Hsueh PR. 2013. Matrix-assisted laser desorption ionization-time of flight mass spectrometry can accurately differentiate between Mycobacterium masilliense (M. abscessus subspecies bolletti) and M. abscessus (sensu stricto). J Clin Microbiol 51:3113–3116. doi: 10.1128/JCM.01239-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Grenfell RC, da Silva Junior AR, Del Negro GM, Munhoz RB, Gimenes VM, Assis DM, Rockstroh AC, Motta AL, Rossi F, Juliano L, Benard G, de Almeida Junior JN. 2016. Identification of Candida haemulonii complex species: use of ClinProTools(TM) to overcome limitations of the Bruker Biotyper(TM), VITEK MS(TM) IVD, and VITEK MS(TM) RUO Databases. Front Microbiol 7:940. doi: 10.3389/fmicb.2016.00940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mari-Almirall M, Cosgaya C, Higgins PG, Van Assche A, Telli M, Huys G, Lievens B, Seifert H, Dijkshoorn L, Roca I, Vila J. 2017. MALDI-TOF/MS identification of species from the Acinetobacter baumannii (Ab) group revisited: inclusion of the novel A. seifertii and A. dijkshoorniae species. Clin Microbiol Infect 23:210 e211–210 e219. doi: 10.1016/j.cmi.2016.11.020. [DOI] [PubMed] [Google Scholar]
  • 18.Zhu B, Xiao D, Zhang H, Zhang Y, Gao Y, Xu L, Lv J, Wang Y, Zhang J, Shao Z. 2013. MALDI-TOF MS distinctly differentiates nontypable Haemophilus influenzae from Haemophilus haemolyticus. PLoS One 8:e56139. doi: 10.1371/journal.pone.0056139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Bruin JP, Kostrzewa M, van der Ende A, Badoux P, Jansen R, Boers SA, Diederen BM. 2014. Identification of Haemophilus influenzae and Haemophilus haemolyticus by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Eur J Clin Microbiol Infect Dis 33:279–284. doi: 10.1007/s10096-013-1958-x. [DOI] [PubMed] [Google Scholar]
  • 20.Hare KM, Binks MJ, Grimwood K, Chang AB, Leach AJ, Smith-Vaughan H. 2012. Culture and PCR detection of Haemophilus influenzae and Haemophilus haemolyticus in Australian indigenous children with bronchiectasis. J Clin Microbiol 50:2444–2445. doi: 10.1128/JCM.00566-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Tille PM. 2012. Bailey & Scott's diagnostic microbiology, 13th ed Mosby Elsevier, St. Louis, MO. [Google Scholar]
  • 22.Yam WC, Yuen KY, Kam SY, Yiu LS, Chan KS, Leung CC, Tam CM, Ho PO, Yew WW, Seto WH, Ho PL. 2006. Diagnostic application of genotypic identification of mycobacteria. J Med Microbiol 55:529–536. doi: 10.1099/jmm.0.46298-0. [DOI] [PubMed] [Google Scholar]
  • 23.Lasko TA, Bhagwat JG, Zou KH, Ohno-Machado L. 2005. The use of receiver operating characteristic curves in biomedical informatics. J Biomed Inform 38:404–415. doi: 10.1016/j.jbi.2005.02.008. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)

RESOURCES