In the present study, we found that nested PCR detected more samples positive for HCMV DNA than qRT-PCR, and nested PCR using PBL DNA had the highest positive rate. Therefore, nested PCR based on PBL DNA is the most sensitive method for detecting HCMV DNA.
False positive caused by cross contamination is a main problem of nested PCR. To avoid such cross contamination, we strictly adopted precautionary measures , such as processing meticulously, introducing negative controls, extracting and amplifying the DNA in different rooms, and changing gloves frequently. Additionally, the positives of nested PCR were confirmed to some extent by the serological assays (Table 4). Furthermore, we performed another nested PCR using a different set of primers and obtained the identical results (data not shown). Therefore, the high sensitivity of this nested PCR in detecting HCMV DNA observed in the study is reliable.
Our findings are in agreement with the previous observation that nested PCR is more sensitive than qRT-PCR [6, 7]. By contrast, several studies [8, 9] reported the opposite results. Additionally, the similar sensitivity or discordant results between qRT-PCR and nested PCR have also been documented by other investigators [10, 20, 21]. One reason for the situation of "different study, different result" may be different patient groups investigated by different studies. Mhiri et al  have suggested that the performance of diagnostic tests may be affected by transplant patients' different immunological responses to different transplant (renal, bone marrow, etc.). However, in the present study, we evaluated all the patients as a whole instead of dividing them into subgroups according to their primary diseases, because we aimed to define a routine test competent enough for most, if not all, patients. Actually, it is impracticable to change the detection for every specific disease during one laboratory's daily work. In addition to the targeted population, other factors such as the PCR primers [23, 24] and the DNA extraction technique  may also influence the amplification efficacy. As Bastien et al  conclude that the PCR is not an isolated technique; instead, it involves a range of techniques, which might affect the outcome depending on a variety of factors. For this study, we believe that more reaction cycles in nested PCR, almost twice of that in qRT-PCR, is the main explanation for the higher sensitivity.
The detection capacity of the qRT-PCR in this investigation appears to be inferior to that of some recorded qRT-PCR [27, 28]. However, a recent multi-center research  has indicated that various HCMV DNA quantitative assays, including both commercial reagents and laboratory-developed assays, are generally not sensitive in detecting viral loads lower than 1000 copies/ml, irrespective of their reported detection limits. Another limitation of this study is that the sample size was relatively small; nevertheless, this study contained prospective serial samples in which HCMV DNA was detected more frequently by nested PCR than by qRT-PCR (Table 3).
Although qRT-PCR has advantages in quantification, automation, and time savings [30, 31], nested PCR is still clinically significant in some settings. Nested PCR is suitable for small number of samples, especially for situations in which quantification is not essential. For examples, detection of HCMV DNA in fetuses or newborns is sufficiently to diagnose an ongoing active infection. Additionally, nested PCR does offer benefits for developing countries because of its low cost.
With appropriate materials adopted in nested PCR, our results demonstrated that positive rate of HCMV DNA in PBL was much higher than that in plasma (34.9% vs. 18.9%, p = 0.002, Table 1), which is in accordance with the reported data [13, 14, 32], although similar sensitivities of plasma and PBL were referred in Ye et al's  and Banan et al's reports . Two reasons may contribute to the higher sensitivity of using PBL DNA: (1) the concentration of DNA extracted from PBL should be much higher than that from 200 μl plasma, and (2) with the biological characteristics of latent infection, HCMV may be harbored in PBL ; when the virus in PBL replicates at a low level, very a few viruses are released into plasma, causing the lower sensitivity of using plasma DNA.
It is considered that positive DNA in serum/plasma is connected more closely with symptomatic infection [15, 16], because serum/plasma DNA reflects disruption of host cells caused by active viral replication. In other words, it seems that PCR using PBL DNA is not suitable for diagnosing active CMV infection, because healthy adults with the latent infection may harbor CMV in their leukocytes. Nevertheless, for immunosuppression patients, the risk of the so-called "latency" to be activated is highly possible. Therefore the early detection of HCMV DNA is especially important for reminding clinicians to be aware of the possible reactivation and to carefully follow-up viral load kinetics. As indicated by relevant studies [31, 32, 36], monitoring HCMV DNA in PBL is helpful in anticipating viral replication. In our study, four patients were negative for HCMV DNA in plasma but positive for DNA in PBL (Table 3), indicating the high sensitivity by nested PCR with PBL. On the other hand, the negative result detected by nested PCR with PBL may be helpful in ruling out HCMV infection, since its negative predictive value is very high (Table 2).