Impact of resistant pneumococcus on rates of acute mastoiditis

METHODS AND MATERIAL

The records of all patients treated at Shands Hospital at the University of Florida, an academic, tertiary-care medical center, with a discharge diagnosis of acute or coalescent mastoiditis between July 1, 1987, and June 30, 1997, were reviewed. Subjects with acute exacerbations of chronic suppurative otitis media and subjects with otitis media and isolated mastoid opacification but no clinical evidence for acute mastoiditis were excluded from the study. Because acute mastoiditis is a clinical diagnosis, patients subsequently found to have cholesteatoma (eg, during surgery or follow-up examinations) were included if they had clinical evidence of acute mastoiditis and no suspicious otologic symptoms (eg, chronic or recurrent otorrhea) or middle ear findings (eg, perforation, scutal erosion, or squamous debris) at presentation. Acute mastoiditis was classified as coalescent or noncoalescent on the basis of the radiographic and intraoperative reports of mastoid bone integrity or erosion. Data collection included patient demographics; causative pathogens isolated from the ear, blood, or spinal fluid; presence of other sequelae; and type of treatment required.

An exact test for directional (1-tailed) linear trend in ordered binomial proportions, available in the EGRET epidemiologic statistical software package (CyTel, Cambridge, MA), was used to determine whether total, pneumococcus, pseudomonas, and staphylococcus acute mastoiditis rates, expressed as proportions of total hospital or otolaryngology admissions, tended to increase over consecutive time intervals. The Pearson χ² test and Fisher exact test were used to test for heterogeneity of admission rates among time intervals. The Fisher exact test was also used to compare the proportion of cases of acute mastoiditis caused by pneumococcus between patients more or less than 2 years of age and between patients from the first and last triennia of the study period. The Spearman rank correlation coefficient was used to assess the degree of rank correlation between time of admission and patient age during the entire study period. The Wilcoxon rank sum test was used to compare age rank sums between patients from the first and last triennia of the study period.

RESULTS

Thirty-four subjects were found to have adequate documentation to support the diagnosis of acute mastoiditis or acute, coalescent mastoiditis. One patient had acute mastoiditis twice, yielding 35 cases of acute mastoiditis during the 10-year study period. Twenty-six cases involved acute mastoiditis without coalescence, and 9 cases involved coalescence. Nine of each of these groups had other complications, ranging from subperiosteal abscess to death (Table 1).

Twenty-two cases (63%) were in males. Forty-one percent involved the right ear. The mean age of subjects at the time of admission was 10.03 years (median 5.1 years, minimum 61 days, and maximum 52.3 years). Thirty-seven percent of subjects were younger than 2 years. Patient age tended to decrease over the course of the study period (Spearman r = −0.45, P = 0.006), with age being significantly lower (P = 0.022) in the last triennia (median age 1.5 years) than in the first (median age 6.9 years, Tables 2 and 3).

The incidence of acute mastoiditis ranged from 1 to 8 cases per year (median 3 and mean 3.6). The absolute yearly rate of acute mastoiditis was quite variable (ie, heterogeneous). To evaluate the rate of mastoiditis with respect to the population at risk, we analyzed the rate of mastoiditis as a function of hospital and otolaryngologic admissions. The rate of hospital admissions rose slightly and reached a plateau during the study period, but the otolaryngologic admissions decreased by one third (Tables 2 and 3). The rate of acute mastoiditis as a proportion of yearly otolaryngologic admissions increased significantly over time (P = 0.024).

A variety of pathogens were isolated from the middle ear or mastoid cultures. Gram-positive cocci were isolated in more than half of the cases (n = 20), and gram-negative bacilli were found in 12 cases (Table 4). Multiple pathogens were isolated from none of the non-coalescent cases, whereas 6 of the coalescent cases were polymicrobic. Gram stains revealed gram-positive cocci in pairs in 5 cases that yielded gram-negative bacilli on culture (Pseudomonas aeruginosa in 2). Another patient had pneumococcus isolated from the blood before admission with mastoiditis caused by pseudomonas and staphylococcus.

OPEN IN VIEWER

Table 1.

Cases of acute mastoiditis and related sequelae

	Acute noncoalescent	Acute coalescent
Total	26	9
Total with complications	9	9
Subperiosteal abscess	5	4
Sigmoid thrombosis	1	4
Facial palsy	2	2
Meningitis	2	2
Otitic hydrocephalus	1	1
Subdural abscess	0	1
Brain abscess	1	0
Death	0	1

Eleven cases of acute mastoiditis (32%) involved pneumococci. Yearly rates of pneumococcal-related acute mastoiditis varied significantly from year to year (P = 0.007 by otolaryngologic admissions, P = 0.065 by hospital admissions). However, pneumococcal-related rates of acute mastoiditis, expressed as a proportion of yearly hospital and otolaryngologic admissions, increased significantly over time (Tables 2 and 3, P = 0.002 and P = 0.002). The proportion of cases of acute mastoiditis caused by pneumococcus was higher during the last 3 years of the study than it was during the first 3 years of the study (56% vs 13%, P = 0.051). Pneumococcal acute mastoiditis tended to occur more often in children younger than 2 years than in subjects older than 2 years (46% vs 23%, P = 0.144). All but 1 strain of pneumococcus (7 of 8) isolated from cases of acute mastoiditis during the past 3 years were penicillin resistant.

P aeruginosa was isolated from 10 cases of acute mastoiditis, all of which were noncoalescent. Pseudomonas-related yearly rates of acute mastoiditis were less heterogeneous (as a percentage of hospital or otolaryngologic admissions, P = 0.078 and P = 0.069, respectively). The rate of acute mastoiditis caused by pseudomonas (as a percentage of hospital or otolaryngologic admissions) did not rise significantly during the study period (Tables 2 and 3, P = 0.081). The tympanic membrane was intact (without perforation, cholesteatoma, or tympanostomy tube) in 4 of 8 cases that yielded P aeruginosa, and specific information on the integrity of the tympanic membrane was reported.

Cholesteatoma, not apparent on examination and without a prior history of chronic infection, was identified after admission in 4 cases of acute mastoiditis. Cholesteatoma was first identified at the time of mastoidectomy (during the hospitalization for acute mastoiditis) in 2 of these cases. In 1 case suspicion of cholesteatoma was raised by CT. The fourth became apparent only after 5 years of follow-up. None of these 4 subjects had mastoiditis caused by pneumococcus: proteus, 2; pseudomonas, 1; and bacteroides, 1.

OPEN IN VIEWER

Table 2.

Rates of acute mastoiditis per 10,000 hospital admissions and median age at admission in 3- to 4-year intervals

Years	No. of admissions	Total (n)	S pneumoniae	P aeruginosa	Staphylococcus aureus	Median age (y)
1988-1991	60,835	1.32(8)	0.16(1)	0.16 (1)	0.16(1)	6.9
1992-1994	92,187	1.19 (11)	0.11(1)	0.65 (6)	0.22 (2)	9.6
1995-1997	73,619	2.17 (16)	1.22 (9)	0.41 (3)	0.41 (3)	1.5
Linear trend		P = 0.093	P = 0.002	P = 0.277	P = 0.189	r = −0.45
Homogeneity		P = 0.243	P = 0.002	P = 0.369	P = 0.645	P = 0.006

OPEN IN VIEWER

Table 3.

Rates of acute mastoiditis per 100 otolaryngology admissions and median age at admission in 3-to 4-year intervals

Years	Admissions	Total (n)	S pneumoniae	P aeruginosa	S aureus	Median age (y)
1988-1991	1531	0.52 (8)	0.07(1)	0.07(1)	0.07(1)	6.9
1992-1994	1929	0.57(11)	0.05 (1)	0.31 (6)	0.10(2)	9.6
1995-1997	1216	1.32 (16)	0.74 (9)	0.25 (3)	0.25 (3)	1.5
Linear trend		P = 0.014	P = 0.0004	P = 0.185	P = 0.156	r = −0.45
Homogeneity		P = 0.028	P = 0.0002	P = 0.308	P = 0.452	P = 0.006

OPEN IN VIEWER

Table 4.

Bacteriology of middle ear or mastoid effusions

Species∗	Noncoalescent	Coalescent	Total
S pneumoniae
Penicillin sensitive	1	0	1
Penicillin resistant	5	2	7
Sensitivities not tested	2	1	3
Viridans streptococci	2	0	2
Streptococcus pyogenes	2	0	2
S aureus	6	0	6
P aeruginosa	10	0	10
Proteus species	0	2	2
Anaerobes	1	1	2
No growth	2	3	5
Not documented	3	0	3

∗Totals may exceed the number of cases because multiple pathogens may have been isolated from each case.

Twenty-one subjects (58%) had documentation of antibiotic therapy being given immediately before admission. Preadmission (before middle ear culture) antibiotic therapy had been received by similar percentages of subjects who had pathogens atypical for acute otitis media (eg, pseudomonas, 64%), no bacterial growth (71%), or resistant strains of primary pathogens (71%). The only pneumococcal isolate known to be penicillin sensitive was recovered from a child who was reported to have received amoxicillin before acute mastoiditis developed.

All subjects were treated with parenteral antibiotic therapy. Twenty percent reportedly had a tympanocentesis, and 29% reportedly had tympanostomy tubes placed as part of their treatments. One third (34%) of subjects underwent a mastoidectomy for evidence of coalescence, development of additional complications, or failure to improve on parenteral therapy. Three subjects underwent a craniotomy to drain brain and subdural abscesses (1 each) and to repair an encephalocele arising through the tegmen mastoideum.

All but 1 patient survived the primary hospitalization. The only fatality occurred in an intravenous drug abuser (HIV negative), who presented in extremis with coalescent mastoiditis, cholesteatoma, sepsis, and meningitis.

DISCUSSION

The impetus for this study was a perception that acute mastoiditis was occurring at a higher than normal rate. Our observations confirmed that the overall rate of acute mastoiditis had increased significantly during the 10-year study period. The only pathogen that was found to have a significantly rising rate of acute mastoiditis was S pneumoniae. All but one of the S pneumoniae isolates that were tested for antibiotic susceptibility were found to be penicillin resistant. These findings suggest that the rising rate of acute mastoiditis is likely the result of emergence of antibiotic-resistant strains of S pneumoniae. As antibiotic susceptibility profiles have been routinely performed on S pneumoniae for only the past few years, this trend is just now being appreciated. Two series of acute mastoiditis that have recently been reported in the US literature have not addressed the significance of pneumococcal antibiotic resistance. 5 , 6

Many factors unrelated to an actual rise in the incidence of acute mastoiditis may have played a role in these increases (ie, artificially elevated the rate). First, our medical center's catchment population (especially for children) may have increased. Because referral bases are highly dynamic entities dependent on the total patient population, third-party payer contracts, and competing clinicians and facilities, it may not be possible to assess the influence of this source of bias. Second, the rates of admissions have decreased for both hospitals and otolaryngology services. Financial pressure from insurance companies and the need to remain fiscally solvent have expanded the number of procedures that are provided on an outpatient basis. Hence, even a static number of cases of acute mastoiditis would yield an increasing rate of mastoiditis as a fraction of hospital admissions. These factors, if they had any role at all, would have led to an increase in the overall numbers of admissions for acute mastoiditis. It is, however, very unlikely that they would have been responsible for the dramatic rise in the rate of pneumococcal disease relative to that caused by other pathogens.

S pneumoniae has long been recognized as a primary pathogen of the human respiratory tract. During the past 40 years, clinicians have grown somewhat complacent about the significance of this pathogen, primarily because it has been so easily eradicated with penicillin. The emergence of pneumococcal antibiotic resistance has been slow to arrive, 4 but it has rekindled interest in its significance in respiratory tract disease. We found that S pneumoniae only recently surpassed P aeruginosa as the pathogen responsible for the majority of cases of acute mastoiditis; however, it may have previously played a greater role in the pathogenesis of complicated otitis media than is appreciated.

Five cases of gram-negative pathogens recovered from culture of the middle ear effusion were initially found to have gram-positive cocci in pairs on the gram stain. A sixth case had pneumococcal sepsis before admission with pseudomonal mastoiditis. All of these cases had received antibiotics before middle ear aspiration, which was known to severely reduce the rate of penicillin-sensitive pneumococcus recovery in acute mastoiditis. 7 These observations raise the possibility that gram-positive cocci (ie, pneumococcus) may have played a role in the pathogenesis of these cases.

Whereas pneumococcus is a known middle ear pathogen, gram-negative bacilli, like pseudomonas species, are opportunistic pathogens, causing disease only where host defenses have been violated. 8 This most commonly occurs in the middle ear as a result of cholesteatoma, tympanic membrane perforation, or a tympanostomy tube. However, pneumococcal infection may render the middle ear more susceptible to pseudomonal infection, especially when host responses to the pneumococcal infection are waning (ie, after appropriate antibiotic therapy). 9 , 10 In contrast, the middle ear is significantly less vulnerable to pseudomonas invasion during active pneumococcal infection. 9 , 10 Hence, pseudomonas would be less likely to superinfect cases of mastoiditis that have been caused by antibiotic-resistant pneumococcus.

These postulates may be of more than casual academic interest. If intermediate levels of pneumococcal antibiotic resistance remain more common than high levels of resistance, oral antibiotics can be expected to help limit the severity of pneumococcal middle ear disease. This would likely lead to continued cases of mastoiditis caused secondary pathogens like P aeruginosa. Most cases of acute mastoiditis caused by pathogens have responded well to medical management alone (eg, none of the pseudomonas cases were coalescent). If higher levels of resistance predominate, more severe forms of pneumococcal disease are likely to be seen. These would be less likely to respond to oral or parenteral antibiotic therapy (ie, short of vancomycin), so surgical therapy may be necessary more often in the future.

The lower age of our patient population, particularly those with pneumococcal mastoiditis, is also noteworthy. Harley et al 6 have recently reported and ascribed this observation to factors such as daycare attendance. Such young children are the least immunologically prepared to battle highly virulent pathogens. As the issue of antibiotic resistance increases, tympanocentesis for middle ear culture may become more valuable and more frequently used in cases of antibiotic treatment failures, so that treatment may be specifically directed. If vigilance is not maintained for potential, serious complications of acute otitis media, these children will be the most likely to have irreversible consequences.

CONCLUSIONS

The incidence of acute mastoiditis has risen significantly during the past 10 years. This appears to be the result of the emergence of pneumococcal antibiotic resistance. Younger children are at an increased risk for pneumococcal mastoiditis. All clinicians should maintain a high level of suspicion for possible intratemporal and intracranial complications resulting from acute pneumococcal otitis media.

We thank Mr Donald Sterner for assisting us with the medical record search, and Ms Amanda Garrigues for administrative support in the preparation of this manuscript.