PCR Can Add to Detection of Pneumococcal Disease in Pneumonic Patients Receiving Antibiotics at Admission

We wish to add to the debate on PCR for the detection of bacteremia in patients on antibiotics (3). We evaluated a nested PCR (nPCR) for Streptococcus pneumoniae DNA in the blood of 86 consecutive adult acute medical admissions in which the admitting intern considered blood culture to be indicated. EDTA blood samples taken at admission were tested by the nPCR, DNA being extracted by the method of Bell et al. (1), using proteinase K digestion and chloroform extraction. The nPCR targeted the pneumolysin gene. The first-round primers (amplifying a region of 501 bp) were 5′-ATGGCAAATAAAGCAGTAAATGACT-3′ and 5′-AGAACCAAACTTGACCTTGAGTTGT-3′. An internal amplification control which coamplified to give a 316-bp product was incorporated. The nested primers amplified a 170-bp region within the first-round amplicon but outside of the amplification control and were 5′-TTTGCCTGGTTTGGCAAGTAGC-3′ and 5′-CCATGCTGTGAGCCGTTATTTTTT-3′. Amplification consisted of 35 cycles for each round, and products were detected by ethidium bromide-stained gels. The nPCR was shown to be specific for S. pneumoniae when tested against a wide range of nonpneumococcal streptococci and unrelated bacterial species. Using serial dilutions of a broth culture of S. pneumoniae NCTC 7466, the first-round amplification detected 7.5 × 10³ CFU and the full nPCR detected 75 CFU.

The nPCR was positive in only one of 42 patients with either nonpneumococcal bacteremia (Escherichia coli in 2 patients and Staphylococcus aureus in 1 patient), localized nonpneumococcal infection, or noninfective conditions. This patient had Streptococcus mitis endocarditis. This apparently anomalous result is, in fact, consistent with recent observations (4). The reaction was inhibited in two samples in this group.

Of three patients with proven bacteremic pneumococcal pneumonia, the nPCR was positive for two and inhibited for one. However, nine nPCR-positive and two inhibited reactions occurred in the remaining 41 patients with negative blood cultures but in whom pneumococcal disease could not confidently be excluded, although it was much more likely in some than in others. Seven positive results occurred in 14 patients presenting with radiological pneumonia but with negative blood cultures compared to only two (one patient with meningitis and one with chronic obstructive pulmonary disease) in 27 patients without pneumonia (P < 0.001). Twelve of these 41 patients were on antibiotics and the nPCR was positive in 7 of these, in contrast to only 2 of the 29 patients not receiving antibiotics (P < 0.001). All seven patients with pneumonia and a positive nPCR result were on antibiotics compared to only three of the seven pneumonic patients with a negative nPCR result. The nPCR results are summarized in Table 1.

TABLE 1.

Results of nPCR for S. pneumoniae DNA for EDTA blood samples from 86 adult medical admissions presenting with possible bacteremia related to clinical assessment category

Clinical assessment category	Total no. of samples	No. of samples with nPCR result
		+ve	−ve	Inhibited
Proven pneumococcal bacteremia	3	2	0	1
Proven nonpneumococcal disease	42	1	39	2
Potentially pneumococcal disease
With pneumonia	14	7	6	1
Without pneumonia	27	2	24	1

Our limited study supports the contention of Heininger et al. that PCR may be useful in bacteremic patients in which antibiotics have prevented cultural diagnosis (2) but also suggests that organism-specific PCRs are most productively targeted at particular syndromes (in this case, an S. pneumoniae PCR for pneumonia patients). Whereas PCR might yet be too expensive for management of individual patients with common conditions such as pneumonia, it will be a valuable additional tool in surveys, the results of which can guide initial empirical therapy.