On 2017 Aug 20, Daniel Weiss commented:

This is a paper which is often cited to tout the robust sensitivity of the Western blot for diagnosis of Lyme disease. However, a careful, critical reading of this paper reveals major flaws. The Western blot is, on the contrary quite insensitive for the diagnosis.

I will limit my comments to study patients without recent erythema migrans. Most workers agree that an erythema migrans rash is diagnostic on its own, and that serologic tests are unnecessary as well as unreliable at this stage of the disease. Dressler performed both a prospective and retrospective study to evaluate two-tier testing. Study subjects were recruited from the authors’ Lyme disease clinic from 1990-1991. To participate they had to live in “an area endemic for Lyme disease”. The G39/40 strain of B. burgdorferi was used to generate antigens for the ELISA assay.

Dressler’s prospective study included 54 highly-selected patients with Lyme disease. The prospective group included patients with Lyme arthritis and those with “inactive” Lyme disease. Many patients had a history of the EM rash. Neuroborreliosis was diagnosed in 39 patients; these patients had CSF pleocytosis, increased CSF protein or EMG evidence of an axonal polyneuropathy not caused by other known disease. Ten of the 39 with neuroborreliosis were declared to be “inactive” although this term was not clearly defined. Importantly, only 21 of the 39 were positive by ELISA—a sensitivity of 54%. Of those cases of “active Lyme disease” who had neuroborreliosis, the sensitivity was 58%.

There were 57 patients with an indeterminate ELISA. One third of those who were thought by the investigators to have “active Lyme disease” had a negative Western blot. Thus, the two-tier test performed poorly even in patients whom expert physicians (Dr. Steere and colleagues) were sure had “active Lyme disease”.

In the retrospective study of Dressler et. two-tiered testing was performed on a highly-selected patient cohort. ELISA was negative in two of 25 with Lyme encephalopathy or polyneuropathy; two more patients’ ELISAs were read as indeterminate. Western blot testing in four of the 25 showed a minimal or absent response. Thus, the sensitivity of the test in these patients chosen because the authors were convinced they had Lyme disease reached only 84%.

On 2017 Aug 20, Daniel Weiss commented:

This is a paper which is often cited to tout the robust sensitivity of the Western blot for diagnosis of Lyme disease. However, a careful, critical reading of this paper reveals major flaws. The Western blot is, on the contrary quite insensitive for the diagnosis.

I will limit my comments to study patients without recent erythema migrans. Most workers agree that an erythema migrans rash is diagnostic on its own, and that serologic tests are unnecessary as well as unreliable at this stage of the disease. Dressler performed both a prospective and retrospective study to evaluate two-tier testing. Study subjects were recruited from the authors’ Lyme disease clinic from 1990-1991. To participate they had to live in “an area endemic for Lyme disease”. The G39/40 strain of B. burgdorferi was used to generate antigens for the ELISA assay.

Dressler’s prospective study included 54 highly-selected patients with Lyme disease. The prospective group included patients with Lyme arthritis and those with “inactive” Lyme disease. Many patients had a history of the EM rash. Neuroborreliosis was diagnosed in 39 patients; these patients had CSF pleocytosis, increased CSF protein or EMG evidence of an axonal polyneuropathy not caused by other known disease. Ten of the 39 with neuroborreliosis were declared to be “inactive” although this term was not clearly defined. Importantly, only 21 of the 39 were positive by ELISA—a sensitivity of 54%. Of those cases of “active Lyme disease” who had neuroborreliosis, the sensitivity was 58%.

There were 57 patients with an indeterminate ELISA. One third of those who were thought by the investigators to have “active Lyme disease” had a negative Western blot. Thus, the two-tier test performed poorly even in patients whom expert physicians (Dr. Steere and colleagues) were sure had “active Lyme disease”.

In the retrospective study of Dressler et. two-tiered testing was performed on a highly-selected patient cohort. ELISA was negative in two of 25 with Lyme encephalopathy or polyneuropathy; two more patients’ ELISAs were read as indeterminate. Western blot testing in four of the 25 showed a minimal or absent response. Thus, the sensitivity of the test in these patients chosen because the authors were convinced they had Lyme disease reached only 84%.