RE: “RACIAL AND ETHNIC DIFFERENCES IN SOCIOECONOMIC POSITION AND RISK OF CHILDHOOD ACUTE LYMPHOBLASTIC LEUKEMIA”

We read with interest the report by Wang et al. (1) on the incidence of childhood acute lymphoblastic leukemia (ALL) by socioeconomic position (SEP) and race/ethnicity. They reported rate ratios comparing the 4 highest levels of SEP with the lowest within Surveillance, Epidemiology, and End Results populations of white, black, Hispanic, and Asian children. While informative, an alternative presentation of these data that was not shown by the authors—incidence rate ratios comparing races/ethnicities within strata of SEP—is also revealing. We calculated these incidence rate ratios (and 95% confidence intervals), with white children as the reference group, using the data provided by Wang et al. (Table 1).

Table 1.

Incidence Rates and Incidence Rate Ratios for Acute Lymphoblastic Leukemia Among Children 0–19 Years of Age, Comparing Nonwhite Races/Ethnicities With White Within Socioeconomic Strata, Surveillance, Epidemiology, and End Results 16 Registries, 2000–2010

SEP	Race/Ethnicity
	White		Black		Black:White		Hispanic		Hispanic:White		Asian		Asian:White
	No.	IR	No.	IR	IRR	95% CI	No.	IR	IRR	95% CI	No.	IR	IRR	95% CI
Lowest	422	2.81	175	1.42	0.51	0.43, 0.61	1,198	4.95	1.76	1.58, 1.97	58	2.24	0.80	0.61, 1.05
Low	656	3.15	130	1.82	0.58	0.48, 0.70	877	4.86	1.54	1.39, 1.71	74	2.32	0.74	0.58, 0.94
Medium	736	3.10	94	1.82	0.59	0.48, 0.73	578	4.36	1.41	1.26, 1.57	119	3.01	0.97	0.8, 1.18
High	892	3.34	69	1.96	0.59	0.46, 0.75	356	3.91	1.17	1.04, 1.32	139	2.93	0.88	0.73, 1.05
Highest	1,121	3.60	42	2.24	0.62	0.46, 0.84	211	3.31	0.92	0.79, 1.07	232	3.65	1.01	0.88, 1.17

Abbreviations: CI, confidence interval; IR, incidence rate; IRR, incidence rate ratio; SEP, socioeconomic position.

At all levels of SEP, the incidence rate ratio was substantially and significantly lower by roughly half among black compared with white children. In contrast, the incidence rate ratio was substantially higher among Hispanic than among non-Hispanic white children in the lowest stratum of SEP, but it fell with rising SEP until becoming null at the highest stratum. Among Asian children there was no significant incidence rate ratio for ALL compared with white children, except for a slightly lower incidence rate ratio in the low-SEP stratum.

While interpretation of surveillance data is necessarily speculative due to the lack of information on individuals, we might draw some tentative conclusions. That the incidence rate ratio for ALL among black compared with white children was lower at all levels of SEP suggests environmental risk is outweighed by substantial African ancestry. Among Hispanic children, the clear gradient of decreasing incidence rate ratio with increasing SEP could represent differences in specific ethnicity but might also reflect increasing European admixture (2). There was a trend towards lower incidence rate ratios for ALL in the 2 lowest strata of SEP among Asian children; we note that Southeast Asian ethnic groups have lower SEP than East Asians in the United States (3).

Collectively these results point to the importance of ancestry in risk of ALL. The results further suggest that among populations with widespread European/non-European admixture, including black and Hispanic children, admixture mapping could identify specific loci and variants conferring differences in risk (4).