Clinical Characteristics of Measles in Previously Vaccinated and Unvaccinated Patients in California

Abstract

Background

Measles vaccine failure was first described in 1972. Over the next 20 years, vaccine failure was extensively studied, but during the last 25 years few investigations have been performed. We describe the clinical characteristics of measles in previously vaccinated and unvaccinated patients in California.

Methods

All confirmed measles cases reported to the California Department of Public Health from 1 January 2000 through 31 December 2015 were reviewed. Clinical characteristics (rates of hospitalization, cough, coryza, conjunctivitis, and fever) were compared between the previously unvaccinated, those who had had 1 dose of vaccine, and those who had had ≥2 doses of measles vaccine.

Results

There were 232 confirmed measles cases in whom vaccination status was verified; 80% were unvaccinated, 9% had had 1 dose of measles vaccine, and 11% had had ≥2 doses of measles vaccine. Subjects who had had ≥2 doses of measles vaccine had lower rates of hospitalization, cough, coryza, conjunctivitis, and fever than subjects who had 1 dose of measles vaccine or who were unimmunized.

Conclusions

Vaccine failure measles cases were less ill than cases that occurred in unvaccinated patients. Nevertheless, these cases still required the same amount of public health effort in tracing contacts as in cases who were unvaccinated.

measles, primary vaccine failure, secondary vaccine failure

(See the Editorial Commentary by Moss on pages 1320–1.)

For >16 years there has been no endemic measles in the United States [1]. However, measles is still endemic and epidemic in Europe, Africa, and Asia and measles importations have occurred frequently during the last decade [2–4]. Many of these importations have resulted in secondary and tertiary cases in nonvaccinated and vaccinated individuals. In the studies by Peter Panum during the 1846 Faroe Island measles epidemic, it was noted that all those who had measles in the previous epidemic in 1781 were protected despite the lack of apparent reexposure for 65 years [5]. During the study of the live Edmonston B measles vaccine during the 1950s and early 1960s, it was found that the postimmunization antibody decay curves were similar to those of children who had had natural measles [6]. Based upon these observations, it was expected that vaccinated children would have lifelong immunity to measles.

Edmonston B measles vaccine was licensed in the United States in 1963 and in 1966 measles vaccines were added to the routine childhood immunization schedule [7–9]. In 1965 and 1968, 2 further attenuated measles vaccines became available in the United States (Schwarz strain and Moraten strain, respectively) [6, 10, 11]. In a 10-year follow-up study of children who had received the further attenuated Schwarz strain measles vaccine and who had had no measles exposure during the 10-year period, Krugman [11] noted that the geometric mean hemagglutination inhibition (HAI) titer was about 1.5 logs to the base 2 lower than the geometric mean titer in children who had received Edmonston B vaccine. In the same study at the 12-year follow-up, it was noted that 9% of the Schwartz strain vaccinees had HAI titers of <1:2. In contrast, all children who had had measles had HAI titers >1:2.

Following the introduction and routine use of measles vaccines in 1966, there was a dramatic decline in measles. However, epidemic measles reoccurred in 1971–1972 [12, 13]. One of us (J. D. C.) was the lead investigator during the 1971–1972 measles epidemic in St Louis. It was noted that unvaccinated children with measles had classic measles and they had a measles immunoglobulin M (IgM) antibody response [12, 13]. A second group of measles patients was identified in which the clinical symptomology was less severe [13]. These modified cases had a typical prodrome (fever ≥38.9°C, cough, and coryza) but the clinical symptomatology was mild, and the rash had the typical distribution and progression of a measles rash but it was not confluent [12]. In these cases, only an immunoglobulin G antibody response was observed [13]. It was unknown whether these modified cases were efficient transmitters of measles virus.

Based on the findings from the St Louis epidemic, it was predicted that a 2-dose vaccination program would be required to achieve the goal of measles elimination in the United States [12]. Since 1972, 2 additional epidemics of measles have occurred in the United States (1976–1977 and 1989–1991) [7]. However, it was not until 1989 that the Advisory Committee on Immunization Practices made the recommendation for a second dose of a measles vaccine [1, 7]. Between the early 1990s and the present, few measles studies have been performed in the United States [8]. Between 2000 and 2015 there were 400 cases of laboratory-confirmed measles in California, and of these cases with a known vaccination history, the vast majority (80%) occurred in unvaccinated individuals; only 20% were vaccine failures. We performed the present study to evaluate the clinical characteristics of illness and epidemiologic aspects of measles in previously vaccinated and unvaccinated persons.

METHODS

All confirmed measles cases reported to the California Department of Public Health (CDPH) from 1 January 2000 through 31 December 2015 were reviewed. Cases were confirmed by either the demonstration of measles-specific IgM antibody or the demonstration of measles RNA by polymerase chain reaction (PCR) in a throat, nasopharyngeal, or urine specimens. Measles-specific IgM antibody studies were requested by the physicians caring for the patients and were performed in commercial laboratories. Measles PCR studies were done in the CDPH laboratory or in commercial laboratories. For all cases, age, sex, immunization dates, exposure history, clinical findings, and transmission history were analyzed. Measles is a reportable disease in California, and data relating to hospitalization, cough, coryza, conjunctivitis, and fever were gleaned from case report forms and supplemented by CDPH and city county/health department personnel. Seventeen cases who received postexposure prophylaxis prior to measles onset were excluded from the analysis as were 151 cases in whom documentation of vaccination or nonvaccination could not be verified. Pairwise P values were computed using Fisher exact test; trend test P values were computed using Cochran-Armitage trend test.

RESULTS

From 2000 through 2015, there were 232 confirmed measles cases with known vaccination status and who had not received postexposure prophylaxis (Table 1). Sixty-nine percent of the cases occurred in 2011–2015 (Table 1). Of the total 232 cases, 186 (80%) were unvaccinated, 20 (9%) had had 1 dose of vaccine, and 26 (11%) had had ≥2 doses of measles vaccine. Virtually all of the 151 cases in whom vaccination status could not be confirmed were adults.

Table 1.

Number of Laboratory-confirmed Measles Cases in California, 2000–2015^a

Year	No. of Measles-containing Vaccine Doses Received
	0	1	≥2	Unknown	Total
2000	5	2	0	10	17
2001	8	2	0	27	37
2002	0	0	0	4	4
2003	2	0	0	3	5
2004	3	0	0	3	6
2005	3	0	0	1	4
2006	3	0	0	3	6
2007	3	1	0	0	4
2008	16	0	0	1	17
2009	5	0	0	2	7
2010	17	1	0	9	27
2011	17	2	1	9	29
2012	4	0	0	4	8
2013	12	1	0	5	18
2014	37	3	11	23	74
2015	51	8	14	47	120
Total	186	20	26	151	383

^aSeventeen cases omitted; 4 received intramuscular immune globulin postexposure prophylaxis (PEP), 2 received measles-mumps-rubella PEP, 8 received first measles-containing vaccine (MCV) at <12 months of age, 3 received MCV within 21 days prior to measles rash onset.

In Table 2, selected clinical characteristics by vaccination status are presented. As can be seen, those with ≥2 doses of vaccine were less likely to be hospitalized or to have cough, coryza, conjunctivitis, and fever than those unvaccinated or those who had only 1 dose of vaccine.

Table 2.

Clinical Characteristics of 232 Laboratory-confirmed Measles Cases in California Analyzed by the Number of Doses of a Measles-containing Vaccine Received^a

Characteristic		0 MCV Doses	1 MCV Dose	≥2 MCV Doses	P Valueb			Trendc
		(n = 186)	(n = 20)	(n = 26)	0 vs 1 Doses	0 vs 2 Doses	1 vs 2 Doses
Hospitalized	Yes	53 (29)	5 (26)	2 (8)	.8017	.0302	.1144	.0428
	No	127 (71)	14 (74)	24 (92)
Cough	Yes	172 (94)	16 (80)	13 (50)	.04616	.0000	.0636	.0000
	No	11 (6)	4 (20)	13 (50)
Coryza	Yes	150 (83)	15 (88)	5 (20)	.7446	<.0001	.00002	<.0001
	No	30 (17)	2 (12)	20 (80)
Conjunctivitis	Yes	123 (71)	9 (53)	6 (24)	.1656	<.0001	.0997	.0000
	No	50 (29)	8 (47)	19 (76)
Fever	Yes	180 (98)	19 (95)	22 (85)	.3417	.00518	.3693	.0032
	No	3 (2)	1 (5)	4 (15)

Data are presented as No. (%) unless otherwise indicated.

Abbreviation: MCV, measles-containing vaccine.

^aMissing hospitalization status on 7 patients; missing cough on 3 patients; missing coryza on 10 patients, missing conjunctivitis on 17 patients, missing fever on 3 patients.

^bFisher exact test.

^cCochran-Armitage trend test.

The time since last measles vaccination in 42 vaccine failure cases is presented in Table 3. The median elapsed time in ≥2 doses of vaccine failure cases was 16.7 years. Of the 26 cases that experienced failure after receiving ≥2 doses of vaccine, 3 were documented to transmit to unimmunized contacts. Case 1 had cough and fever but no coryza or conjunctivitis and transmitted to a close friend who spent >5 hours at the patient’s house. Case 2 had mild illness but transmitted to 2 family members, and case 3 had cough and fever but no coryza or conjunctivitis and transmitted to a household contact.

DISCUSSION

We have done the largest study in the vaccine era on the clinical characteristics of measles in unvaccinated and vaccinated people. It is noted that illness in the previously unvaccinated is the most severe and a trend analysis notes that those who received ≥2 doses of vaccine were less ill than those who were 1-dose vaccine failures (Table 2). Primary and secondary measles vaccine failures were first described in Cincinnati and St Louis in 1972–1973 in relation to the return of epidemic measles following its initial control by vaccination [12–16]. In the St Louis studies in which 1 of us (J. D. C.) was the lead investigator, it was noted that in primary vaccine failure the host initially had an IgM immunologic response to measles virus whereas in secondary vaccine failure, the host had immunologic memory for measles virus [12, 13].

The original studies on primary and secondary measles vaccine failure in St Louis and at other locations during the 1971–1972 epidemic were compromised by several factors [12–16]. At that time, measles vaccine was often not stored properly, and it was often administered with immune globulin at various concentrations. For secondary vaccine failure it was known that the vaccinees had immunologic memory for the virus, but it was not known if they actually had measurable antibody titers. On the laboratory side, the routine technique at that time for measuring IgM antibody was by the reduction of titer by treating the serum with 2-mercapto ethanol. This method lacked sensitivity. The alternative technique (density gradient separation) was labor intensive and could only be done in the research setting [12, 13].

Another problem at that time was that antibody values were most often measured by HAI. This technique was insensitive. The plaque reduction neutralization (PRN) test was later developed by Albrecht et al; it was 60 times more sensitive than HAI [17]. Enzyme-linked immunosorbent assays, which are routine today, are also more sensitive than HAI [18].

In 1985 Chen et al had the opportunity to study a measles outbreak on a university campus [19]. Since there had been a Red Cross blood drive on the campus about 1 month prior to the measles outbreak, they were able to look at preexposure measles antibody values in cases and noncases. There were 9 subjects who had preexposure PRN measles titers of ≤120 IU. Of this group, 8 were measles vaccine failures. In contrast, none of the 71 subjects who had measles PRN titers >120 IU experienced measles.

In 1989, Markowitz et al [20] did a complete review of all articles between 1972 and 1989 on measles vaccine failure. They noted that persons who experienced an initial antigenic stimulation from either natural measles or measles vaccine generally exhibited an anamnestic, or secondary immune, response after either revaccination or exposure to natural measles. They also noted that secondary vaccine failures had been documented and that the illness in those with secondary vaccine failure was normally mild. They also noted that there was a more rapid decline in measles antibody in populations that were isolated and did not get boosts in antibody from exposure to wild virus compared with populations where measles was prevalent.

Only 3 of our 26 (12%) vaccine failure cases with ≥2 doses were the source of infection in contacts, and all 3 contacts who developed infection had significant contact with the cases and no contact with community cases, either as household contacts or close friends. In recent years a number of investigators have studied measles outbreaks and looked at primary and secondary vaccine failures [21–28]. In general, secondary vaccine failure cases were less ill than cases that occurred in unvaccinated individuals. Also, in contrast with our present findings, transmission from secondary vaccine failure cases to those exposed has not been reported.

Although it is reassuring that transmission from secondary vaccine failure cases is rare, we do not know whether this lack of transmission is the result of good public health procedures or due to a decreased level of contagion. Good public health procedures include aggressive case finding; the identification of contacts; and, in potentially susceptible persons, the institution of postexposure administration of vaccine or immune globulin. Also, in a number of the published studies with no transmissions, we do not know the actual numbers of those susceptible among the exposed persons.

It should be noted that all of our vaccine failure cases and all of the vaccine failure cases in the other studies required the same amount of public health effort in tracing contacts as would occur in primary measles cases. Also, since measles rash illness in secondary vaccine failure cases can occur without cough, coryza, or conjunctivitis, investigators should do measles PCR studies more frequently.

Our study has several limitations. We do not have proof (based upon the lack of an IgM antibody response or on avidity studies) that those who received ≥2 doses of vaccine are in fact secondary failures, as opposed to cases who failed to develop primary measles immunity after both vaccine doses. Nevertheless, the statistically milder clinical presentation of these cases suggests a clinical process, which would be most consistent with illness associated with waning, but not absent, immunity. We have also assumed that our 20 cases of measles in subjects who had received 1 dose of a measles-containing vaccine were primary vaccine failures. The finding of decreased cough frequency (80% vs 94%) in 1-dose vaccine failure subjects compared with illness in unvaccinated subjects, as well as the overall trend, suggests that some of the 1-dose vaccine failure cases are secondary and not primary vaccine failures.

In 1978, 1 of us (J. D. C.) and colleagues noted that the antibody response following revaccination in subjects with secondary failure cases was not long lasting [29]. A more recent study in Wisconsin also noted that revaccinated subjects had low measles antibody titers 5 years after revaccination [30].

As noted in Table 3, the median time between last vaccination and vaccine failure in ≥2-dose failures was 16.7 years and in 1-dose failures it was 21 years. Because of lack of exposure during the last 17 years and resulting antibody titer boosting, it is possible that some middle-aged persons in the United States, if exposed, are likely to develop measles due to waning immunity. As measles is endemic and epidemic in Asia, Africa, and Europe, middle-aged US travelers with measles antibody that has waned may be at risk of infection [31]. In the same vein, importations are a yearly occurrence in the United States [3, 4].

Table 3.

Time (Years) Since Last Measles Vaccination to Rash Onset Date in 42 Vaccine Failure Cases

No. of Vaccine Doses	No. of Cases	Mean (Years Since Last Vaccination)	Median (Years Since Last Vaccination)	Range
≥2 dosesa	23	15.4	16.7	6.0–23.6 years
1 doseb	19	18.2	21	52 days–49.2 years

^aThree patients had missing dates for last vaccination.

^bOne patient had missing date for last vaccination.

From the previous studies referred to above [30, 31], we know that boosting within 10 years of immunization does not result in a lasting response. However, we do not know whether revaccination of persons >20 years since immunization will be more successful. In our opinion, this should be studied in a cohort of middle-aged persons.

Notes

Acknowledgments. We appreciate the statistical help of Jeffrey A. Gornbien and guidance by Jennifer Zipprich, Kathleen Harriman, and Kathleen Winter of the Immunization Branch of the California Department of Public Health.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Presented in part: American Epidemiological Society Meeting, 2015, Berkeley, CA; IDWeek, 2015, San Diego, CA.