Originalia M. Maiwald, C. Stockinger, D. Hassler, M. von Knebet Doeberitz, H.-G. Sonntag Evaluation of the Detection of Borrelia burgdorferi DNA in Urine Samples by Polymerase Chain Reaction Summary: It is difficult in some cases to identify an infection caused by Borrelia burgdorf- eri and to monitor the effect of therapy. Seropositivity will persist even after successful treatment and therefore may suggest ongoing infection. For direct detection orB. burgdorf- eri DNA in human urine samples, the polymerase chain reaction (PCR) was evaluated. A published primer system was selected, which amplifies a 259 bp fragment from the gene en- coding the 23S rRNA. The lower detection limit of the primer system was 10 fg of extract- ed It. burgdorferi DNA. Several methods for the pretreatment of urine samples were test- ed. Of these, the Geneclean® kit (Bio 101, USA) showed the best results. A total of 114 urine samples from 74 patients belonging to three clinical groups was investigated: (i) 51 samples from 26 patients with active Lyme disease, (ii) 36 samples from 27 patients with previous infection but no symptoms at the time the urine was collected, and (iii) 27 samples from 21 seronegative control patients without Lyme disease. B. burgdorferi DNA was de- tected in 25 urine samples of 17 patients with active disease, whereas 26 samples from this group of patients were negative. Only one asymptomatic case with previous infection showed a positive result, and the urine samples of the patients without Lym e disease were uniformly negative. Two of four patients from whom samples before and directly after on- set of therapy were available converted from negative to positive PCR results after initia- tion of therapy, accompanied by the symptoms of a Jarisch-Herxheimer reaction. It can be concluded from these results that a positive PCR from urine is with high probability an in- dicator of active Lyme disease. On the other hand, as only 17 of the 26 patients with active infection were positive, a negative PCR result does not exclude active infection. Introduction Since the discover)' of Borrelia burgdorferi as the etiolog- ic agent of Lyme disease [1], Lyme borreliosis has attract- ed increasing attention both in Europe and in the United States. Clinical symptoms of Lyme disease are variable, but generally, early and late manifestations can be differ- entiated. By analogy with syphilis, the division into three stages has been widely- accepted [2]. Symptoms of stage one are the characteristic erythema migrans as well as non- specific systemic complaints, the second stage is character- ised by neurologic and cardiac sequelae, and stage three includes chronic skin, neurological and arthritic symp- toms. Most important for diagnosis are clinical history and examination. Because microscopic or cultural detection of Borrelia burgdorferi are difficult and unreliable, serology is the most commonly applied tool of laboratory diagno- sis. However, serology for B. burgdorferi has some inher- ent problems, such as the delayed rise of antibody titres in early stages of Lyme disease, and in some instances, the occurrence of false positive results due to the presence of cross-reacting epitopes with other bacteria [2,3]. A clear limitation of serology for B. burgdorferi is the dis- tinction between current and previous infection. This issue becomes critical when the effect of therapy must be mon- itored, and when Lyme disease must be distinguished from illnesses with similar symptoms, especially in endemic re- gions, where the prevalence of positive antibody titres is high. After successful therapy, it takes at least several months before a significant decrease of antibody titres can be observed. Non-specific symptoms, such as chronic fa- tigue, may persist in some patients despite effective thera- py. IgM antibodies, which are often a valuable aid in oth- er infectious diseases, are not helpful for the monitoring of therapy in Lyme borreliosis, because only a fraction of the patients with active disease exhibit positive IgM titres, es- pecially in the late stage of Lyme borreliosis [4]. The culture of B. burgdotferi is generally considered to be difficult and has a low yield from samples from infected patients [2]. Therefore, it is desirable to have a method for the direct detection of the pathogen in patient specimens that is more sensitive than culture. The polymerase chain reaction (PCR) has been proposed as such a method. It is very sensitive, detecting as few as 5-10 spirochetes in one reaction, and it has been applied to ticks [5], biopsies from rodents [6], as well as cerebrospinal fluid [7,8], blood sam- ples [9], urine samples [10,11], and skin biopsies [12,13] from patients. However, the practical value of the PCR as Received: 14 March 1994/Revision accepted: 13 March 1995 Dr. reed. M. Maiwald, C Stockinger, Prof. Dr. reed. H.-G. Sonntag, Hy- giene Institut der Universit~it, Im Nenenheimer Feld 324, D-69120 Hei- delberg; Dr. reed. D. Hassler, Allgemeinarztpraxis, Untere Hofstatt 1-3, D-76703 Kraichtal; PD Dr. med. M. von Knebel Doeberitz, Projekt- gruppe Angewandte Tumorvirologie, Deutsches Krebsforschungszen- trum, Im Neuenheimer Feld 242, D-69120 Heidelberg, Germany. Infection 23 (1995) No. 3 © MMV Medizin Verlag GmbH Mttnchen, Mt~nchen 1995 173 / 47