Identifying a Level of Neutralizing Antibody That Correlates With Protection From Clinical Mumps Disease During a 2017 Mumps Outbreak Among Military Service Members

Abstract

Background

In 2017, a mumps outbreak occurred in a US military barracks. Serum collected at service entry was used to compare pre-exposure with presumptive vaccine-induced antibody levels from persons who developed mumps (cases) and potentially exposed persons who did not develop mumps (non-cases). Sufficient information to determine levels of exposure during the outbreak was not available.

Methods

Pre-outbreak serum samples from the Department of Defense Serum Repository were available from 254 potentially exposed service members. Twelve developed clinical symptoms and had post-outbreak serum collected. All sera were tested with a mumps-specific enzyme immunoassay for immunoglobulin M, immunoglobulin G (IgG), and IgG avidity. The neutralizing antibodies to vaccine strain (Jeryl Lynn [JL], genotype A) and wildtype virus (genotype G) was assessed by a plaque reduction neutralization test. A Fisher exact test and receiver operator characteristic curve were used to analyze the antibody response for non-cases and mumps cases.

Results

Eight mumps cases were laboratory confirmed. Pre-outbreak neutralizing antibody titers to JL and genotype G mumps virus and pre-outbreak IgG index values were proportionately lower for most cases as compared with exposed non-cases. When compared with potentially exposed non-cases, cases with clinical symptoms had greater odds of having a pre-outbreak JL titer <41 and a genotype G titer <16.

Conclusions

We identified potential correlates of protection for mumps neutralizing antibody titers against JL and genotype G mumps viruses.

Military recruits live and train in close quarters that are favorable for the transmission of respiratory viruses such as mumps. Prior to the introduction of the mumps vaccine in the late 1960s, there were frequent mumps outbreaks among the military that caused significant disruptions to military operations. These outbreaks decreased with the introduction of the 2-dose measles, mumps, and rubella (MMR) vaccine schedule, which became universally administered to recruits in 1991 [1–3].

In 2010, the Department of Defense discontinued the practice of universal MMR vaccination for new recruits, instead targeting individuals based on vaccination history and screening for measles and rubella antibody titers [4]. Relatively low numbers of mumps cases and outbreaks have been reported in the subsequent years [3].

In 2017, the United States was in the midst of a 4-year mumps resurgence, with cases and outbreaks in 49 of 52 US jurisdictions, mostly among fully vaccinated individuals [5]. One mumps outbreak occurred on an army base in Texas, coinciding with outbreaks throughout the United States. The index case at the Texas base (case 1) transferred into Advanced Individual Training in February after completing Basic Military Training in another state. On 25 February, he experienced bilateral mandibular and parotid swelling, was hospitalized, and tested indeterminate for mumps immunoglobulin M (IgM). This specimen was sent to the Centers for Disease Control and Prevention (CDC) for confirmatory testing and was positive according to the CDC IgM capture assay, which is a more sensitive and specific assay with limited availability [6]. The index case recalled a trainee with whom he had been housed in his previous Basic Military Training class, who had been diagnosed with pharyngitis with facial swelling in January prior to his transfer to Advanced Individual Training. Neither of these trainees was tested for mumps at the Basic Military Training site, according to personal correspondence with Brooke Army Medical Center medical staff [7].

During January to April 2017, asymptomatic cases and symptomatic cases may have been missed since the index case was identified retrospectively. Between 7 April and 4 May 2017, 11 Advanced Individual Training trainees from this company of 270 personnel presented to Brooke Army Medical Center with symptoms of mumps. An outbreak investigation was performed by Army Public Health Nursing, and 252 individuals in the company were given a third dose of MMR vaccine per the guidance of the CDC and Advisory Committee on Immunization Practices for mumps outbreak control [7, 8]. A military training company offered a unique opportunity to study a mumps outbreak, providing a controlled congregate setting and the opportunity of testing pre- and post-infection serum. The Department of Defense Serum Repository houses serum specimens collected from active and reserve component armed forces members for medical surveillance, linked to personnel and medical records in the Defense Medical Surveillance System. All trainees have sera collected prior to initial entry training, pre- and post-deployment, and roughly every 2 years for required HIV screening [9]. Pre-outbreak sera from mumps cases and non-cases during this outbreak provided an opportunity to assess the relationship between levels of pre-outbreak vaccine-induced neutralizing antibody and protection from mumps.

METHODS

Department of Defense Serum Repository Samples

Trainees (N = 254) from this outbreak had multiple repository specimens collected. However, only the most recent specimen was included for analysis; collection dates ranged from June 2004 to January 2017, with a median specimen collection of 198 days before the first case was identified on 25 February 2017. In the original outbreak investigation, 249 trainees were investigated; yet, an additional 5 trainees were included in the outbreak investigation by the CDC after contact tracing [6]. All 254 trainees either provided documentation of 2 childhood doses of MMR vaccine or had screening for measles and rubella immunoglobulin G (IgG) performed soon after entry into the service, with screening data obtained from the Medical Protection System and the Defense Enrollment Eligibility Reporting System. Of the 254 trainees, 57 were either negative or equivocal for measles and rubella IgG at entry into service and received a dose of MMR vaccine. This study was given a non-research determination by the CDC and the Department of Defense Institutional Review Board.

Laboratory-Confirmed Mumps Cases

Laboratory-confirmed mumps cases were those with a positive IgM test result by CDC IgM capture assay and/or positive reverse-transcription polymerase chain reaction (RT-PCR) with classic mumps symptoms such as parotid/salivary gland swelling and orchitis.

Non-cases

The degree of exposure that each trainee had to a mumps case was difficult to assess due to the congregate setting, with shared facilities for sleeping, eating, and training; thus, all trainees who were not identified as mumps cases were classified as non-cases. Non-cases were either asymptomatic individuals or those with prodromal symptoms (low-grade fever, headache, myalgia, malaise, and anorexia) who had a negative test result by CDC IgM capture assay or RT-PCR and a positive test result for an alternate etiology at the Brooke Army Medical Center.

Laboratory Testing at the CDC

Reference repository sera for all 254 trainees were tested with the following assays, and post-outbreak sera collected from the mumps outbreak investigation (12 trainees) were sent to the CDC for confirmatory testing and to assess antibody response after initial laboratory testing was conducted by the military hospital [7].

Plaque Reduction Neutralization Tests

The plaque reduction neutralization assays to genotype A (vaccine strain, Jeryl Lynn [JL]) and genotype G (wild type strain) were used as previously described to assess neutralization titers to both genotypes. Plaque reduction neutralization end point titers were determined to be the highest dilution of serum that gave ≥50% plaque reduction per the Karber formula [10].

Enzyme-Linked Immunosorbent Assay Testing

Reference repository sera and sera collected from symptomatic cases were tested for mumps-specific IgM via the CDC mumps IgM capture enzyme immunoassay (EIA) as previously described. The CDC mumps IgM capture is the most sensitive serologic method, detecting 46% to 71% of RT-PCR–confirmed cases [6]. Mumps-specific IgG was measured with a commercially available indirect enzyme-linked immunosorbent assay kit (Zeus Scientific). Positive, negative, and equivocal categorization of sera was determined through the cutoff values specified by the manufacturer based on an index value/OD ratio. EIA index values were defined as follows: ≤0.90, seronegative; 0.91 to 1.09, equivocal; and ≥1.10, seropositive.

IgG Avidity

IgG avidity determinations were made for cases, symptomatic non-cases, and a subset of repository serum specimens from exposed non-cases by using an end point titration assay as previously described [11]. Results were interpreted as low avidity if the end titer avidity index was ≤30% and high avidity if >30% [11].

Statistical Analyses

Data analysis and visualization were performed with R version 4.3.0 and Prism version 8.0 (GraphPad). To analyze the data, all pre-outbreak specimens were divided into quartiles based on their JL neutralizing titer to show the distribution of antibody levels across non-cases and mumps cases. Statistical significance between quartiles and mumps cases was defined as P < .05. The analysis to compare cutoff points for JL and genotype G to distinguish cases and non-cases was described by Cortese et al using the Fisher exact test and odds ratio calculated in R version 4.3.0 [12]. The cutoffs determined by this analysis predicted the best cutoffs associated with the level of exposure that resulted in a mumps case. These cutoffs were further characterized by receiver operator characteristic (ROC) curve analysis to compare the area under the curve for the neutralization titers with genotypes A and G. The ROC analysis established cutoffs to identify mumps cases by distinguishing the differences between the neutralization response for repository serum and post-outbreak specimens.

RESULTS

Laboratory-Confirmed Mumps Cases and Non-cases

Twelve trainees were symptomatic and exhibited typical mumps symptoms (parotid/salivary gland swelling, orchitis, fever, malaise [7, 13]. Of the 12 trainees, 8 tested positive by RT-PCR or the CDC IgM capture EIA (Table 1). Of the 5 RT-PCR–positive cases, 4 (80%) tested IgM positive and 1 was IgM indeterminate. Seven (88%) patients exhibited parotitis, 4 (50%) reported fever, 2 (25%) reported sore throat, and 1 (14%) reported orchitis.

Table 1.

Clinical and Diagnostic Results for Symptomatic Service Members During a 2017 Military Mumps Outbreak

				Date			Titer		Result
No.a	Age,b y; Sex	CDC Laboratory Classification (Case)c	Other Testing and Alternative Diagnosis	Sample Collection	Symptom Onset (Days to Collection)	Outbreak Dose of MMRd	Jeryl Lynn	Genotype G	CDC IgM	Zeus IgG	Zeus Index Value	Avidity (%)	PCR Result
1	…; M			22 Nov 2016e			33.9	6.8	–	+	2.50	Low (20)
		Mumps (P)	+ entero, + flu	1 Mar 2017f	25 Feb 2017 (4)	No	539.5	163.7	+	+	5.47	High (60)	NR
2	19; M			28 Aug 2016e			12.4	7.4	–	EQV	0.92	Not tested
		Mumps (P)	+ rhino, + entero	28 Apr 2017f	7 Apr 2017 (21)	Yes	978.2	848.0	+	+	5.01	High (62)	–
3	20; M			31 May 2016e			27.8	7.0	–	+	2.64	High (65)
				5 Oct 2016e			21.3	8.3	–	+	2.41	High (67)
		Mumps (P)	+ C pneumoniae	28 Apr 2017f	8 Apr 2017 (20)	Yes	386.5	405.7	+	+	5.93	QNS	–
4	19; M			7 Jul 2016e			7.0	4.5	–	EQV	1.06	QNS
		Mumps (C)	+ CMV IgM/IgG	2 May 2017f	10 Apr 2017 (22)	Yes	309.5	179.2	+	+	5.43	High (725)	+
5	23; M			26 Mar 2012e			36.0	14.7	–	+	2.47	High (43)
				14 Aug 2016e			29.5	9.9	–	+	2.08	High (55)
				22 Nov 2016e			32.7	11.6	–	+	2.58	High (36)
		Mumps (C)	+ CMV IgG	28 Apr 2017f	11 Apr 2017 (17)	Yes	3570.4	2530.0	+	+	5.91	QNS	+
6	25; M			29 Jul 2016e			54.0	23.0	–	+	1.27	High (72)
				23 Nov 2016e			42.9	17.1	–	+	1.24	UND
		Mumps (C)	+ rhino, + entero	28 Apr 2017f	11 Apr 2017 (17)	No	1410.8	792.0	+	+	5.09	High (53)	+
7	19; M			28 Jun 2016e			167.4	55.2	IND	+	3.30	High (90)
		Not a case	− rhino, − monospot	26 Apr 2017	21 Apr 2017 (5)	Yes	176.3	31.5	–	+	3.15	High (63)	–
8	20; F			10 Aug 2016e			8.6	8.6	–	–	0.74	Not tested
				22 Nov 2016e			6.2	5.5	–	–	0.80	Not tested
		Mumps (C)	− rhino, monospot	26 Apr 2017f	23 Apr 2017 (3)	Yes	30.6	10.1	IND	+	2.84	High (51)	+
9	19; F			20 Oct 2016e			33.9	19.1	–	+	3.30	High (64)
		Not a case	− rhino	26 Apr 2017f	24 Apr 2017 (2)	Yes	119.1	41.6	–	+	3.23	QNS	–
10	20; M			15 Jun 2016e			301.5	50.4	–	+	3.16	High (65)
		Not a case	+ C pneumoniae	25 Apr 2017f	25 Apr 2017 (1)	Yes	217.4	40.6	–	+	3.21	High (59)	–
11	23; M			14 Oct 2016e			33.5	5.7	–	+	1.10	UND
		Mumps (C)	− rhino	2 May 2017f	1 May 2017 (1)	Yes	178.7	84.5	+	+	5.72	High (62)	+
12	25; M			5 Oct 2016e			52.9	19.1	–	EQV	0.97	Not tested
		Not a case	+ rhino, + entero	4 May 2017f	4 May 2017 (0)	Yes	54.3	20.5	–	+	1.33	High (51)	–

Patients are ordered by onset date and CDC classification as a mumps case (bold).

Abbreviations: +, positive; –, negative; C pneumoniae, Chlamydia pneumoniae; CDC, Centers for Disease Control and Prevention; CMV, cytomegalovirus; entero, Enterovirus; EQV, equivocal; F, female; IgG, immunoglobulin G; IgM, immunoglobulin M; IND, indeterminate; M, male; MMR, measles, mumps, and rubella; monospot, mononucleosis spot test; NR, no result; PCR, polymerase chain reaction; PRN, plaque neutralization titer; QNS, quantity not sufficient (for testing); rhino, Rhinovirus; UND, undetermined.

^aWith the exception of patient 1, patients were in the same unit.

^bAge indicates age at onset.

^cConfirmed case (C): positive result on reverse transcriptase PCR test for mumps-specific nucleic acid. Probable case (P): diagnosed by the Department of Defense Public Health and meets clinical criteria for parotid/salivary gland swelling with prodromal symptoms and epidemiologic linkage or supportive laboratory evidence (positive IgM result).

^d20 April 2017.

^eRepository collection date.

^fCollection after onset of symptoms.

Serum from the index case (case 1) was retested by the CDC capture assay and was mumps IgM positive. A buccal specimen for RT-PCR was not collected since the patient was identified later through public health contact tracing [7]. The patient had a low-avidity mumps IgG result for the repository serum and a high-avidity mumps IgG result for the outbreak sample. The outbreak sample was collected 4 days after symptom onset and 3.5 months after the repository sample, indicating IgG maturation as previously described [11].

All 8 laboratory-confirmed mumps cases had a ≥4-fold rise in neutralizing antibody titers to either or JL and genotype G. The EIA index value also increased >2-fold between pre- and post-outbreak samples. Besides the index case, the remaining 7 mumps cases had high-avidity IgG to mumps, ranging from 36% to 90% for 3 of 7 pre-outbreak samples and 5 of 7 post-outbreak samples (Table 1).

The 4 remaining symptomatic trainees were determined to be non-cases, having alternate etiologies identified by serologic testing. They showed no increase in neutralizing titer to either mumps genotype and no increase in EIA index values for samples collected pre- and post-outbreak (Table 1).

Neutralizing Antibody Titers to JL and Genotype G Among Cases and Non-cases

The geometric mean titer (GMT) for the 8 mumps cases increased for JL and genotype G: >24-fold for JL with a pre-outbreak GMT of 19.2 and >33-fold for genotype G with a pre-outbreak GMT of 7.6. However, the GMT for the symptomatic non-cases stayed relatively constant: for JL pre- and post-outbreak titers of 97.5 and 125.5 and for genotype G pre- and post-outbreak titers of 31.7 and 32.3 (Table 2). For all non-cases, the pre-outbreak GMTs for JL (74.7) and genotype G (19.6) were significantly higher than the pre-outbreak titers for mumps cases (P = .0011 and P = .0164, respectively). Quartiles for pre-outbreak non-cases were stratified by their JL neutralization titer, the pre-outbreak neutralization titers to JL and genotype G were similarly low for mumps cases and quartile 1 (Q1) non-cases. But for cases, the post-outbreak titers for JL (GMT, 451.9) and genotype G (GMT, 253.9) were significantly higher than all the pre-outbreak quartiles for both genotypes (Supplementary Table 1).

Table 2.

EIA Results and Neutralizing Antibody Titers to Jeryl Lynn and Genotype G Pre-outreak and Post-outbreak

	Jeryl Lynn Neutralization Antibody Titer			Genotype G Neutralization Antibody Titer			EIA Index Value
	Pre-outbreak	Post-outbreak	Fold Difference	Pre-outbreak	Post-outbreak	Fold Difference	Pre-outbreak	Post-outbreak
Mumps cases
Patient 1	33.9	539.5	15.9	6.8	163.7	24.1	2.50	5.47
Patient 2	12.4	978.2	78.9	7.4	848.0	114.7	0.92	5.01
Patient 3	21.3	386.5	18.1	8.3	405.7	48.9	2.41	5.93
Patient 4	7.0	309.5	44.2	4.5	179.2	39.8	1.06	5.43
Patient 5	32.7	3570.4	109.2	11.6	2530.0	218.1	2.58	5.91
Patient 6	42.9	1410.8	32.9	17.1	792.0	46.3	1.24	5.09
Patient 8	6.2	30.6	4.93	5.5	10.1	1.8	0.80	2.84
Patient 11	33.5	178.7	5.33	5.7	84.5	14.8	1.10	5.72
GMT (95% CI)
Mumps cases (n = 8)	19.2 (10.2–36.3)	451.9 (135.8–1504.0)	23.57	7.6 (5.3–10.99)	253.9 (61.4–1050)	33.2	1.58	5.18
Symptomatic non-cases (n = 4)	97.5 (19.6–486.1)	125.5 (47.2–332.0)	1.3	31.7 (12.5–80.9)	32.3 (19.2–54.5)	0.98	2.68	2.73
Repository Sera for non-cases (n = 242)	74.4 (65.1–84.9)	…	…	19.33 (17.2–21.7)	…	…	3.09	…

Abbreviations: EIA, enzyme immunoassay; GMT, geometric mean titer.

EIA Results Among Cases vs Non-cases

A significant increase in EIA index values was observed between pre- (1.58) and post-outbreak (5.18) sera for mumps cases, while the EIA index values for symptomatic non-cases remained constant at 2.70 (Table 2). The pre-outbreak average EIA index values for all symptomatic and non-symptomatic non-cases (3.09) was significantly higher (P = .0247) than that for cases (1.58; Supplementary Table 1).

Avidity Results Among Non-cases

Repository specimens were divided into quartiles based on neutralization titers against JL to ensure that avidity testing was done proportionally on a subset of all samples. Avidity was determined for a subset of 25 randomly chosen specimens from each quartile of non-cases, with the lowest quartile having titers <40 and the highest having >180. Of the 100 samples tested, 95 had high-avidity IgG and 5 had low-avidity IgG. When samples with low avidity were excluded, mean end titer avidity index by quartile was as follows: Q1, 53% (95% CI, 44%–58%); Q2, 57% (95% CI, 53%–61%); Q3, 66% (95% CI, 60%–71%); Q4, 63% (95% CI, 57%–68%; Supplementary Figure 1).

Defining a Correlate of Immunity

To establish a correlate of protection based on mumps-specific neutralizing antibodies, pre-outbreak neutralization titers against JL and genotype G from cases and non-cases were compared. As compared with non-cases, mumps cases for JL (7/8) had 16.7-times greater odds (P = .0014) of having a pre-outbreak JL titer <41, while mumps cases for genotype G (7/8) had 9.49-greater odds (P = .023) of having a pre-outbreak genotype G titer <16 (Table 3). Mumps cases had JL titers ranging from 6 to 43 and genotype G titers from 4 to 17. Pre-outbreak titers for JL and genotype G were plotted with the cut-offs established by the odds ratios and encompassed 7 of 8 mumps cases (Figure 1).

Table 3.

Comparison of Mumps Neutralization Titers for Cases and Non-cases

	% (No.)
Strain: Titer	Cases	Non-cases	Odds Ratio (95% CI)	P Valuea
Jeryl Lynn
<16	37.5 (3)	6.10 (15)	9.06 (1.29–52.00)	.0135*
<19	37.5 (3)	9.75 (24)	5.48 (.8–30.26)	.0419*
<31	50.0 (4)	19.5 (48)	4.09 (.73–22.82)	.0577
<41	87.5 (7)	29.3 (72)	16.73 (2.09–763.59)	.0014**
<64	100 (8)	45.1 (111)	Infinity (2.01–infinity)	.0020**
<83	100 (8)	55.3 (136)	Infinity (1.34–infinity)	.0109*
Genotype G
<7	50.00 (4)	14.2 (35)	5.96 (1.06–33.57)	.0212*
<8	62.50 (5)	18.7 (46)	7.17 (1.34–47.75)	.0094**
<16	87.50 (7)	42.3 (104)	9.49 (1.19–432.8)	.023*
<21	100 (8)	52.8 (130)	Infinity (1.48–infinity)	.0086**
<34	100 (8)	69.9 (172)	Infinity (.71–infinity)	.1094

Abbreviation: EIA, enzyme immunoassay.

^aComparison of proportions below each titer cut-off by Fisher exact test.

*P < .05. **P < .01.

Figure 1.

Neutralizing antibody response for repository sera (pre-outbreak) by quartiles (Q1–Q4) and mumps cases. For each trainee, the neutralizing response to genotype A (Jeryl Lynn) is on the y-axis and genotype G on the x-axis. Mumps cases are depicted by closed circles. Q1 is depicted by open circles, Q2 by open squares, and Q3/Q4 by upward/downward facing triangles. Dotted lines show the cutoffs based on the odds ratios (Table 3): 41 for Jeryl Lynn and 16 for genotype G. Of the 8 mumps cases, 7 are within these established cutoffs.

ROC analysis comparing post-outbreak neutralization titers for confirmed cases against pre-outbreak neutralizing titers for non-cases was conducted for both genotypes to show the significant differences between pre- and post-outbreak titers that resulted in a mumps case. ROC analysis resulted in areas under the curve of 0.851 (95% CI, .6707–1.00; P = .0007) for JL and 0.899 (95% CI, .7353–1.000; P = .0001) for genotype G (Table 4). Correlation between the neutralization antibody responses to both genotypes for each mumps case was significant, with an r² of 0.519 (data not shown).

Table 4.

Percentages of Mumps Cases and Non-cases for Each Cut-off Established by the ROC Curve

	ROC Analysis: Jeryl Lynn		ROC Analysis: Genotype G
	Titer >41	Titer >301	Titer >16	Titer >54
AUC (P value)	0.851 (.0007)		0.899 (.0001)
Sensitivity/specificity, %	87.5/28.9	75.0/91.4	87.5/41.87	87.5/87.76
% (No.)
Mumps cases post-exposure (n = 8)	100 (8)	75 (6)	87.5 (7)	87.5 (7)
Repository mumps cases (n = 8)	12.5 (1)	0	12.5 (1)	0
Repository non-cases (n = 246)	70.7 (174)	8.9 (22)	57.7 (142)	13 (32)

Abbreviation: AUC, area under the curve; ROC, receiver operator characteristic.

ROC analysis showed that a JL neutralization titer cutoff >41 had a sensitivity and specificity of 87.5% and 28.9% for distinguishing between cases and non-cases. All mumps cases had post-outbreak JL titers >41, but only 12.5% (1/8) had pre-outbreak titers >41, with a titer of 42.9. The percentage of pre-outbreak titers >41 for the non-cases was 70.7% (174/246). The sensitivity (75.0%) and specificity (91.4%) for a JL titer >301 was the most accurate cutoff for identifying a mumps case by plaque reduction neutralization. Seventy-five percent (6/8) of the post-outbreak mumps cases had titers >301, while none of the pre-outbreak titers for mumps cases were close to this cut-off. Of the pre-outbreak non-cases, only 8.9% (22/246) had titers >301 (Table 4).

ROC analysis based on the genotype G neutralization cutoff >16 had a sensitivity and specificity of 87.5% and 41.87% for distinguishing cases and non-cases. Of the 8 confirmed mumps cases, 87.5% (7/8) had post-outbreak titers >16. Only 12.5% (1/8) had pre-outbreak titers >16, with a titer of 17.1. The percentage of pre-outbreak titers >16 for the non-cases was 78.1% (143/246). A genotype G titer >54 was the cut-off for identifying a mumps case by plaque reduction neutralization, the sensitivity and specificity were 87.5% and 87.7%. Almost all cases (87.5%, 7/8) had post-outbreak titers >54, with 1 case having a titer of 10. None of the pre-outbreak titers for mumps cases had titers >54, while 13% (32/246) of the non-cases had pre-outbreak titers >54 (Table 4).

DISCUSSION

The recent resurgence of mumps in the United States is likely due to several factors, including a generally weak initial immune response to vaccine, waning immunity after vaccination, and lower neutralizing antibody titers to circulating wild type strains as compared with the vaccine strain [14–16]. There is currently no accepted correlate of protection for mumps. This mumps outbreak provided an opportunity to identify a correlate of protection by comparing pre- and post-outbreak neutralizing antibody titers. Identifying a protective level of neutralizing antibody would be useful in seroprevalence studies to identify susceptible age groups for intervention and to evaluate the potential usefulness of administering a third MMR dose. Additionally, due to the inherently weak immune response generated by the current mumps vaccine, development of new mumps vaccines targeted at the currently circulating genotypes is under consideration [10, 14–16]. A correlate of protection will be critical in evaluating the immunogenicity and efficacy of novel mumps vaccine candidates.

Determining a protective neutralizing antibody level for mumps has been challenging due to logistical factors (ie, studies are dependent on access to serum samples from patients and contacts prior to an outbreak) and subsequent collection of appropriately timed post-outbreak samples before mumps outbreak guidance is implemented [12, 17]. In our study, 252 of 254 trainees received a third dose of MMR for outbreak control, and suspected cases were immediately isolated, limiting our ability to assess post-exposure titers attributed to clinical exposure vs a response to a third dose and likely limiting the scope of the outbreak [18, 19].

Mumps virus can cause recurrent disease in some individuals; therefore, it cannot reasonably be expected that an attenuated vaccine will provide complete protection for all vaccinated persons. Additionally, mumps can cause asymptomatic infection in approximately ∼20% of unvaccinated persons; estimates of asymptomatic infection among vaccinated persons range from 20% to 66% [17, 20–23]. Possible asymptomatic infections among exposed vaccinated persons and the administration of a third dose of MMR in this outbreak make identifying these cases more complicated. Persons exposed may have a neutralization titer that protects against clinical disease but not infection, further complicating the establishment of a protective cutoff between cases and non-cases.

Incomplete IgG avidity maturation may increase susceptibility to mumps infection. Previous studies have shown mumps antibody avidity to be significantly lower than for measles in some vaccinated persons [11]. In our study, similar levels of avidity maturation were observed between most pre-outbreak samples collected from cases (end titer avidity index, 36%–67%) and non-cases (31%–69%). The low avidity result for the index case indicates that this patient was undergoing a primary immune response at the time of the repository collection due to an earlier mumps virus infection; the index case did not receive an MMR dose prior to this collection. Data on the time course of IgG avidity maturation in individuals who are infected is extremely limited but suggest that maturation to high IgG avidity occurs approximately 160 to 180 days after onset of parotid symptoms [24]. Overall, it is unclear how the rate of IgG avidity maturation contributes to mumps susceptibility.

Additionally, pre-outbreak neutralization titers against JL and genotype G were significantly lower for symptomatic cases as compared with the non-cases in our study. Only 8 cases were identified in this outbreak, even with Q1 non-cases having similar levels of neutralizing antibody to both genotypes; this suggests that the level of immunity needed to confer protection from mumps is a continuum dependent on not only the existing level of neutralizing antibody but also a cellular response at the time of exposure and the duration and intensity of exposure [25]. Unfortunately, we do not have detailed information to determine the level of exposure for non-symptomatic cases in Q1, nor do we have post-outbreak sera for these non-cases to determine whether their mumps neutralization titers increased during the outbreak.

A study by Cortese et al investigating the correlate of immunity for mumps was inconclusive, but similar titers seen in our study were observed by the same assays to characterize plasma specimens from a blood drive that preceded a mumps outbreak [12]. The authors reported that proportionately more case patients than nonexposed persons had a pre-outbreak JL titer <31 (64% vs 27%), a genotype G titer <8 (55% vs 14%), and an EIA index value <1.4 (64% vs 9%) [12]. Our comparable results suggest that a JL titer >41 and a genotype G titer >16 could be used as a correlate of protection. Titers from this study and ours overlapped, and no distinguishable cutoff separated all mumps case from non-cases [12, 17]. Reproducibility is more challenging at low levels in neutralization assays, so a cutoff titer >16 for genotype G may be more reasonable. Further studies could provide additional insight, but it is unlikely that any study will identify a distinct cutoff between cases and non-cases due to asymptomatic infections.

Our study is subject to limitations that may have affected our ability to distinguish between those who were exposed and not exposed. Some cases may have been missed due to underreporting by the trainees that were housed with the confirmed cases, as trainees may not have wanted to lose training time and possibly not graduate with their company, a known issue in military training settings [26]. In addition, administration of a third MMR dose occurred during the outbreak, and though important for outbreak control, it did limit our ability to obtain specimens from exposed soldiers to assess post-exposure titers and identify asymptomatic infections. Third, many trainees in Q1 had low antibody levels and could have been asymptomatic or not infected. The assumption that all service members were potentially exposed was necessary because individual exposure levels could not be determined, although this was likely an overestimate.

In the Latner et al study, at 1 month after the third-dose vaccination, a 3-fold increase in JL neutralization titers was observed, while the symptomatic cases in our study who received a third dose had a 5- to 109-fold difference in titers to JL [27]. Based on the approximate correlate of protection of 41, this lowest quartile could be vulnerable to infection after 1 year, suggesting that a third MMR dose does not provide long-term protection but can help provide a short-term boost during an outbreak [27, 28]. In our study, 8 laboratory-confirmed cases had pre-outbreak titers below or very near the cutoffs of 41 for JL and 16 for genotype G, which would encompass the 3 laboratory-confirmed cases identified by Cortese et al [12]. These results suggest that symptomatic cases do not have protective levels of antibodies to JL and genotype G prior to mumps infection. The neutralizing antibody response to most infections typically peaks 2 to 4 weeks after infection [29, 30]. The 8 post-outbreak neutralization titers to both genotypes had a median collection date of 3 weeks after onset of symptoms. ROC analysis showed that cases had post-outbreak titers to both genotypes, >301 for JL and >54 for genotype G, with only a small percentage of non-cases having titers above these cutoffs, indicating that these increases were due to infection. These large increases correlate with symptomatic mumps cases that had titers below the cutoffs that we suggested in this study. The identification of these cutoffs as a potential correlate of protection for mumps may be useful indicators for predicting protection from mumps disease and inform studies to develop improved mumps vaccines [15, 16].

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank the members of the CDC Mumps Team for their contributions to this manuscript: Stephen N. Crooke, Jaime Tappe, Elizabeth Krow-Lucal, and Mariel Marlow. We are also appreciative to the Brooke Army Medical Center Outbreak Investigation Team and Defense Centers for Public Health for their part in this investigation: Laura Pacha, Cynthia Perry, John Ambrose, and Kelly Gibson.

Author contributions. S. B. S., N. S. C., and C. J. H. contributed to the conception of the project, study design, data collection and analysis, interpretation of data, and drafting of the manuscript. S. M., L. N., H. C., N. N. J., C. C. B., N. B. M., and A. E. M. contributed to data collection, laboratory procedures, interpretation of data, and review of the manuscript.

Patient consent. This study was given a nonresearch determination by the CDC and the Department of Defense Institutional Review Board.

Financial support. This work was supported by the Centers for Disease Control and Prevention (contract number D-762-17).