Diagnosis of Mycoplasma pneumoniae Pneumonia with Measurement of Specific Antibody-Secreting Cells.

Mycoplasma pneumoniae (Mp) is reported to be the most common bacterial cause of community-acquired pneumonia (CAP) in hospitalized U.S. children (1). However, current diagnostic tests, including PCR of upper respiratory tract (URT) specimens and serology, do not differentiate between Mp infection and carriage (2). Mp carriage in the URT is found in up to 56% of healthy children (2, 3). A ≥4-fold increase in IgG levels is still used in most centers to confirm Mp infection but has low sensitivity (4) and is not helpful in acute clinical management (3). In the absence of an accurate diagnostic test, it is not surprising that studies and meta-analyses on the efficacy of antibiotics are inconclusive for Mp CAP in children (5, 6).

Circulating antibody-secreting cell (ASC) responses have been demonstrated to be more rapid and shorter-lived than antibody responses (7). We hypothesized that Mp-IgM-ASCs circulate in peripheral blood only for a few days or weeks after Mp infection, whereas Mp-DNA in the URT and serum antibodies persist for months. We aimed to evaluate the measurement of Mp-IgM-ASCs by enzyme-linked immunospot (ELISpot) assay as a new test for diagnosing Mp CAP.

Methods

Pediatric patients with CAP (n = 152) and control subjects (n = 156) were enrolled from May 2016 to April 2017 after written informed consent. Inclusion criteria for patients with CAP were clinical diagnosis of pneumonia (fever >38.5°C and tachypnea [8]) in previously healthy children aged 3–18 years. Children <3 years were excluded because of a high probability of viral coexistence in the URT (8). Control individuals included healthy children (undergoing elective surgical procedures) and siblings of patients with CAP (with higher chance of being asymptomatic carriers) without recent (≤1 wk) respiratory tract infections.

In all enrolled children, pharyngeal swabs were taken for Mp real-time PCR (9). If additional consent was given, blood samples also were collected in control individuals and patients with CAP (before antibiotic treatment) to test for the presence of Mp-IgM-ASCs by ELISpot assay (detailed in the legend of Figure 1) (10) and Mp-IgM, Mp-IgG, and Mp-IgA by ELISA (2). Finally, we only included children with fresh (isolated ≤4 h) peripheral blood mononuclear cells to avoid poor ELISpot assay performance resulting from decreased ASC viability (in case of isolation >4 h after sampling) or reduced ASC recovery (after a freeze–thaw cycle) (10). Samples and clinical data (using a standardized questionnaire) were collected at follow-up visits at <2 weeks, 2 weeks to 2 months, and 2–6 months.

Figure 1.

(A–C) Comparison of diagnostic test results between patients with community-acquired pneumonia (CAP) (n = 63; median age, 6.0 yr; interquartile range [IQR], 4.4–10.2 yr) and control subjects (n = 21; median age, 6.1 yr; IQR, 4.9–7.9 yr). CAP samples were collected at disease presentation with a median of 12 days after onset of symptoms (IQR, 11–16; range, 2–29). PCR-positive patients with CAP testing negative for Mycoplasma pneumoniae (Mp)-IgM–antibody-secreting cells (ASCs) are indicated in black. Differences in medians are shown with the corresponding P value (Mann-Whitney U test). (A) Mp-DNA levels in pharyngeal swab samples. (B) Mp-IgM levels. The dashed line represents the cutoff for the test (17 U/ml), with a lower limit of quantification of 5 U/ml. (C) Mp-IgM-ASC responses. (D–F) Mp-IgM-ASC enzyme-linked immunospot (ELISpot) assay. Assays were performed as described previously (10) and were specific for the following antigens: Mp (detergent extract enriched for highly specific adhesion protein P1, 2 µg/ml; Virion/Serion), influenza A and B virus (FluarixTetra quadrivalent influenza virus vaccine, 6 µg/ml; GlaxoSmithKline), and total IgM (affinity-purified antibodies to human immunoglobulin light chains λ and κ as positive control, 10 µg/ml; Southern Biotech). The negative control consisted of phosphate-buffered saline (PBS) only in uncoated wells. Representative patterns of ELISpot wells with 10,000 peripheral blood mononuclear cells (PBMCs) per well are shown. Spots were counted by an ELISpot reader (AID) using predefined settings. The spots identified by the machine were manually inspected for the presence of artifacts. Antigen-specific spot counts were calculated as the mean of three wells minus the mean number of spots in PBS wells. Data were expressed as ASCs per 106 PBMCs (10). Corresponding chest X-rays of patients with CAP are shown on the right. The pulmonary infiltrate is indicated with a frame. (D) Mp PCR-positive CAP. (E) Mp PCR-negative CAP. (F) Mp PCR-positive healthy control (carrier). Notably, although the applied protocol has a rather long overall turnaround time (∼24 h), alternative protocols were developed recently that suggest more rapid (∼6–8 h) ASC detection (10). Optimizing such protocols in the future may help translate the Mp-IgM-ASC ELISpot assay into routine clinical care. HC = healthy control.

Assuming that 15% of pairs switch from PCR to IgM-ASC ELISpot assay (positive to negative) and 2% from IgM-ASC ELISpot assay to PCR (negative to positive), we calculated a sample size of 85 children (patients and control subjects) to achieve 80% power and 5% 2-sided significance. Dichotomous data were reported as percentages and compared with χ2 or Fisher’s exact test. P values are two tailed with significance at <0.05 (R software environment, version 3.4.0).

Results

Mp-DNA was detected by PCR in 29% (n = 44/152) of patients with CAP and 8% (n = 12/156) of control individuals (P < 0.001). We were able to perform a complete diagnostic work-up for Mp in 63 patients with CAP and 21 control individuals (n = 12 elective surgery; n = 9 siblings), which included the Mp-IgM-ASC ELISpot assay of fresh peripheral blood mononuclear cells and Mp-IgM ELISA from serum samples. Chest X-rays were routinely performed in 60 (95%) of 63 included patients with CAP, and 98% (n = 59/60) met the World Health Organization criteria for radiological pneumonia.

In the CAP series, Mp-DNA was detected by PCR in 32 (51%) patients, 29 (46%) of whom showed positive responses in the Mp-IgM-ASC ELISpot assay (P = 0.722; Figure 1). In the three Mp PCR-positive patients with CAP who tested negative for Mp-IgM-ASCs, another pathogen was found based on the results of multiplex PCR from pharyngeal swab samples and specific serology (Table 1). All patients who were Mp PCR positive and Mp-IgM-ASC positive were also Mp-IgM seropositive, but Mp-IgM was also found in 3 (10%) patients with CAP who tested negative by Mp PCR and Mp-IgM-ASC ELISpot assay.

Table 1.

Diagnosis of Patients with CAP Who Are PCR Positive for Mp but Negative by Mp-IgM-ASC ELISpot Assay

	Patient 1			Patient 2	Patient 3
Demographic characteristics
Age, yr	4.5			5.9	3.4
Sex	M			M	M
Microbiological characteristics
PCR
Mp-genomic copy numbers/ml*	415			213	177
Other pathogens detected by multiplex PCR†	Adenovirus			Rhinovirus	RSV A
	Rhinovirus				Human bocavirus
	Chlamydophila pneumoniae
Serology‡
Time point of serum sample collection after onset of symptoms, d	1	7	20	12	19	40
Mp-specific antibodies	Negative	Negative	Negative	Negative	Negative	Negative
IgM, ≤17 U/ml	5	13	12	<5	<5	<5
IgG, ≤15 U/ml	<3	3	<3	<3	<3	<3
IgA, ≤14 U/ml	<2	<2	<2	<2	<2	<2
C. pneumoniae–specific antibodies	Negative	Negative	Negative	—	—	—
IgM, <10 U/ml	5	8	9	—	—	—
IgG, <10 U/ml	<4	<4	<4	—	—	—
Adenovirus-specific antibodies	Positive	Positive	Positive	—	—	—
IgM, <1 Index	<1	<1	<1	—	—	—
IgG, <13 U/ml	19	23	24	—	—	—
RSV-specific antibodies	—	—	—	—	Positive	Positive
IgM, <1 Index	—	—	—	—	<1	<1
IgG, <15 U/ml	—	—	—	—	27	16
Diagnosis	Adenovirus			Rhinovirus	RSV A

Definition of abbreviations: ASC = antibody-secreting cell; CAP = community-acquired pneumonia; Mp = Mycoplasma pneumoniae; RSV = respiratory syncytial virus.

Bold indicates the summary and conclusion of all testing in the table.

All three patients with CAP had significantly lower pharyngeal Mp-DNA levels than Mp-IgM-ASC ELISpot-positive patients with CAP (Figure 1).

The multiplex PCR FTD Respiratory pathogens 21 (FTD21) assay (Fast-track Diagnostics) was used to test for respiratory pathogens other than Mp in these three patients. Notably, we are unable to provide information on cocolonization or coinfection in other patients with CAP and control individuals, as we did not systematically test for other pathogens. However, Mp was recently shown to frequently coexist with other bacterial and viral pathogens in the upper respiratory tract of both symptomatic and asymptomatic children (1, 2). Therefore, detection of other pathogens would likely not have changed the conclusions of this study.

Serum samples were tested with Serion ELISA classic tests (Virion/Serion). No serological assay was available for rhinovirus. It is important to note that reinfections are often characterized by weak or absent specific IgM antibody responses (3, 8).

Pharyngeal swab and blood samples were collected at inclusion (n = 84) and follow-up visits (n = 52, 41 patients with CAP and 11 control individuals) and resulted in more than two visits in 42 (81%) and more than three visits in 27 (52%) children, performed at <2 weeks (n = 43), 2 weeks to 2 months (n = 38), and 2–6 months (n = 38). In contrast to Mp-IgM-ASCs, which were found only within 6 weeks after symptom onset, Mp-DNA and/or Mp-IgM persisted ≥4 months in 7 (11%) patients with CAP. Only 10 (34%) Mp-IgM-ASC–positive patients showed a ≥4-fold increase in Mp-IgG, whereas the remaining (n = 19, 66%) had significantly increased Mp-IgG already in first serum samples (median, 49 U/ml; range, 20–125 U/ml; cutoff, 15 U/ml), making a ≥4-fold increase very unlikely.

Among control subjects, Mp-DNA was detected by PCR in 10 (48%) children. All of these tested negative for Mp-IgM-ASCs (P < 0.001; Figure 1). Six (29%) control individuals had positive Mp-IgM, of whom 1 (5%) showed a ≥4-fold increase in Mp-IgG at follow-up. Although 4 (19%) control subjects were serologically or PCR positive for up to 2 months, Mp-IgM-ASC responses were undetectable during 6-month follow-up.

Discussion

In this longitudinal observational study, the measurement of Mp-IgM-ASCs by ELISpot assay allowed a differentiation between infection and carriage. We detected Mp-IgM-ASCs as early as 2 days after symptom onset, with a peak at presentation of CAP at median 12 days. Another previous study about Mp-IgM-ASCs in 12 Mp-seropositive children with CAP corroborated these findings (11). The inclusion of asymptomatic carriers in our study was essential to assess the usefulness of Mp-IgM-ASC detection as a diagnostic test that can distinguish between carriage and infection.

In the absence of a “gold standard” for Mp infection diagnosis, the discriminative potential of the Mp-IgM-ASC ELISpot assay could not be quantified by measures of diagnostic accuracy such as sensitivity and specificity (12). In fact, if a unanimously accepted reference standard is lacking, alternative study designs, as the longitudinal observational study design chosen in this study, may be more appropriate than test accuracy studies to determine the benefit of a new diagnostic test (12). However, it is important to note that our study population represents a convenience sample from a hypothesis-generating single-center study with small control group and longitudinal follow-up in only two-thirds of the children, at pragmatically arranged visits instead of standardized weekly follow-ups. We thus cannot rule out that unintended selection bias occurred. A larger confirmatory study is needed, now that the potential for the Mp-IgM-ASC ELISpot assay has been shown. Improving the early diagnosis of Mp infection in patients with CAP by the Mp-IgM-ASC ELISpot assay may help future interventional studies assessing the effect of antimicrobial treatment in the management of Mp CAP (5, 6).