Errors and omissions in the study of snuff use and hypertension

Dear Sir, We are writing to point out that the study of high blood pressure and hypertension amongst Swedish male snuff users by Hergens et al. [1] contains several apparent errors. In addition, the study omits prevalence estimates for the 37546 subjects who had an Inpatient Register diagnosis of hypertension prior to baseline, and it omits follow-up information on hypertension for 35464 subjects who were healthy at baseline but did not have repeated measurements. In Table 1 (‘baseline-cohort’ column) and Table 2 (all rows except ‘≥65’), the number of ever snuff users is larger than the sum of the numbers of former and current users (the largest discrepancy of 57 is in the ‘baseline cohort’ column in Table 1). Similarly, the number of all workers in the ‘repeated-measurements’ column of Table 1 (n = 42005) is larger than the sum of never- and ever-snuff users, and the percentages for the highest two dose categories, 5% and 3%, are incorrect. We also point out that the results in this manuscript differ from those in Table 10 of the original thesis of this work published by the Karolinska Institute [2]. There is an ambiguity in Tables 2 and 3, which report age-specific odds ratios (ORs) but contain footnotes describing age adjustment. It is highly unusual for age-adjustment to be carried out within 5-year age intervals, so the authors should either explain this logistic model or correct the footnote. Hergens et al. [1] present detailed information about the 5915 workers with high blood pressure at baseline, giving prevalence and OR estimates according to snuff use, age and consumption level (Tables 1–3). But the authors provide no corresponding estimates for the 37546 workers who had an Inpatient Register diagnosis of hypertension prior to baseline. As defined in this study, high blood pressure and hypertension are two separate case definitions for the same condition. Failure to provide the effect estimates for hypertension is incomprehensible, and the study must be considered incomplete until this important omission is corrected. Hergens et al. [1] emphasize in the Abstract, Introduction and Discussion, the longitudinal nature of their study. However, the person-time contributed by cohort members is not mentioned anywhere in the manuscript. In addition, longitudinal data are presented only in Table 4, which also omits critical outcome data. The table lists relative risks (RRs) for an Inpatient Register diagnosis of hypertension during follow-up amongst all workers who were healthy at baseline (n = 77469), and amongst a subset of workers who had repeated measurements (n = 42005), but there is no RR information for a second subset of 35464 workers who were healthy at baseline and did not have repeated measurements. The RRs for the repeated-measurement subset, regardless of snuff use (i.e. ever, former, current and all consumption levels) are higher than those for all workers, indicating that the former may have had other characteristics that contributed to these elevated risks. Hergens et al. [1] must provide RRs for hypertension amongst the 35464 workers who were healthy at baseline and did not have repeated measurements. Otherwise, the authors’ aim of assessing ‘...the risk of… hypertension amongst male long-term users of snuff, particularly based on longitudinal data’, is not achieved. Hergens et al. concluded that their ‘results are of potential public health importance as the prevalence of snuff use is high in Sweden and that hypertension is one of the major risk factors for cardiovascular disease’. The potential public health importance of this study is contingent upon the resolution of the errors and omissions that we have described. Our research is supported by unrestricted grants from smokeless tobacco manufacturers to the University of Louisville (US Smokeless Tobacco Company and Swedish Match AB) and to the University of Alberta (USSTC). The terms of the grants assure that the grantors are unaware of this letter, and thus had no scientific input or other influence with respect to its design, analysis, interpretation or preparation. Neither of us has any financial or other personal relationship with regard to the grantors.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.