. 2013 Feb 14;9(2):e1003174. doi: 10.1371/journal.ppat.1003174

Table 4. Correlation of PCR-based assays with viral outgrowth assay.

Patient populations^a:	Total				Acute				Chronic
Assay	r^b	95% CI^c	P^d	n	r^b	95% CI^c	P^d	n	r^b	95% CI^c	P^d	n
Droplet digital PCR for HIV-1 DNA in PBMC	0.20	−0.17 to 0.52	0.29	30	0.46	0.23 to 0.85	0.18	10	−0.038	−0.47 to 0.41	0.87	20
Droplet digital PCR for HIV-1 DNA in resting CD4⁺ T cells	0.45	−0.06 to 0.77	0.08	16					0.10	−0.53 to 0.66	0.76	11
Alu PCR for integrated HIV-1 DNA in PBMC	0.70	0.36 to 0.88	0.0008	19	0.25	−0.62 to 0.84	0.59	7	0.76	0.34 to 0.93	0.0038	12
Alu PCR for integrated HIV-1 DNA in resting CD4⁺ T cells	0.41	−0.13 to 0.76	0.13	15					0.38	−0.33 to 0.81	0.28	10
qPCR for HIV-1 DNA in rectal biopsy CD4⁺ T cells	0.26	−0.22 to 0.64	0.28	19					0.17	−0.35 to 0.62	0.53	16
qPCR for HIV-1 RNA in rectal CD4⁺ T cells	0.46	−0.01 to 0.76	0.053	18					0.38	−0.16 to 0.75	0.38	15
RNA/DNA ratio in rectal CD4⁺ T cells	0.57	0.14 to 0.82	0.013	18					0.52	0.01 to 0.81	0.047	15
Droplet digital PCR for 2-LTR circles in PBMC	0.19^e	−0.18 to 0.52	0.31	30	0.12^e	−0.55 to 0.69	0.75	10	0.06^e	−0.39 to 0.49	0.81	20
Droplet digital PCR for 2-LTR circles in resting CD4⁺ T cells	0.38^e	−0.15 to 0.74	0.15	16					0.037^e	−0.57 to 0.62	0.91	11
Single copy PCR assay on plasma virus	0.070^f	−0.30 to 0.42	0.71	30	−0.14^f	−0.71 to 0.54	0.71	10	0.14^f	−0.32 to 0.55	0.55	20

Correlations were analyzed for the entire study population (n = 30) and for subpopulations of study subjects starting HAART during acute/early (n = 10) or chronic (n = 20) HIV-1 infection. Because of limitations in the size of blood or tissue samples, not all assays were performed for all subjects, and the n for each assay is shown in the last column. Correlation coefficients were only computed for n>5.

Correlation coefficient for correlation with the viral outgrowth assay. Except where indicated, the Pearson correlation coefficient is shown.

95% confidence intervals for the correlation coefficient. Strong correlations were considered to be formally excluded when the upper limit of the 95% confidence interval was <0.6.

P values <0.05 are indicated in bold type.

Because of the large number of samples in which 2-LTR circles were below the limit of detection (2 copies/10⁶ cells), a Spearman's rank correlation coefficient was computed using a value of 1 copies/10⁶ cells for the samples in which 2 LTR circles were not detected.

Because of the large number of plasma samples that had HIV-1 RNA levels below the limit of detection (0.2 copies/ml), a Spearman's rank correlation coefficient was computed using a value of 0.1 copies/ml for negative assays.