The Impact of Season and Time Served in Remand Prisons on the Development of Tuberculosis

  Published: 29.12.06 Updated: 30.12.2006 10:58:14

T. Lobacheva1, T. Asikainen 2,3, J. Giesecke 1,2
1 Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden, 2 Department of Epidemiology, Swedish Institute for Infectious Disease Control, Stockholm, Sweden, 3Department of Mathematical Statistics, Stockholm University, Stockholm, Sweden

The incidence of tuberculosis (TB) in Russia increased almost three-fold from 1991 to 2000 when it reached 90.7 per 100 000 population. Many efforts have been made to improve the situations in the prison system since 2000. The Russian Federal Program “Prevention and Control of Social Diseases for 2002-2006” includes a section entitled “Urgent Measures of Tuberculosis Control in Russia” that addresses the TB situation. After 2000, the incidence stabilised and decreased to 83.2/100 000 in 2003 (1).

According to official data, the TB incidence in Saint Petersburg, Russia was 44/100 000 in 2002 and 40/100 000 in 2003 (2), approximately half the national incidence. The health situation in the Russian prison system has received considerable attention due to the high TB incidence (3, 4). During the last ten years, the overall TB incidence in the Russian prison system increased from 860/100 000 in 1992 to 2,028/100 000 in 2002 (5). The highest incidence was seen in 1997 when 4 060 cases per 100 000 were reported (6). The TB incidence in the prison system in Saint Petersburg and the surrounding region (Leningrad oblast) was slightly lower compared with the TB incidence in the whole penitentiary system in Russia; 1983 cases per 100 000 were reported in 2002 (7).

All detainees are held in custody at remand prisons (SIZOs) while awaiting trial. The maximum duration of time served in SIZOs was officially shortened to 6 months from summer 2002 (7). However, some prisoners are held in detention for more than 2 years while waiting for trial due to delays in the Court of Justice.
Saint Petersburg has four SIZOs and 2 additional SIZOs are situated in the Leningrad oblast. Two of the SIZOs for male prisoners in Saint Petersburg have medical divisions for detainees with TB.

The TB incidence in Saint Petersburg’s remand prisons was more than 20 times higher than in the city during 2002 (7). Comparison with previous years is difficult because the method for calculating incidence was revised in 2004 and there was a considerable decrease in the total number of prisoners due to an amnesty. However, the incidence was still high in 2004 with SIZO-1 in Saint Petersburg reporting 1 053 cases per 100 000 (8).

It has been estimated that more than 50% of the total population in Russia belongs to a risk group for contracting TB (9). Together with unhealthy work and living conditions, addictive habits (smoking, alcohol- and drug-abuse) and a low income, being incarcerated at a correction center is one of the most important  factors for contracting pulmonary TB (10, 11). Conditions in remand prisons are difficult. Factors such as staying in overcrowded and sparsely ventilated cells, spending limited time outside, long detention time, etc., all contribute to the transmission and development of different diseases including tuberculosis (1).

 Seasonal variation of TB incidence has been studied in different countries. TB incidence is higher between April and June in India (12), during summer and autumn in Spain (13) and during summer in the United Kingdom (14, 15). In contrast, most cases are detected during the winter period (January) in Russia (16).

The main objective of this study was to analyze the risk of developing TB in a remand prison as a function of time since detention and to investigate the impact of seasonality with respect to TB incidence.

Material and Methods

The study was conducted in Saint Petersburg in the two SIZOs for men that have TB divisions (SIZO-1 and SIZO-4). In order to estimate the risk of developing TB as a function of time in detention, data was collected on the distribution of detention time for all prisoners. These data were retrieved from the SIZO archives that contain information on all prisoners who have been released from remand prison during a particular year.

A total of 6 500 detainees were randomly sampled (17% of the average prison population in SIZO-1 and 25% in SIZO-4 during 3 years) using the follow sampling method. All prisoners from the archives in SIZO-1 with surnames beginning with the randomly chosen letters “A”, “B” and “M” who had been released in 2000, 2001 and 2002 were included. From SIZO-4 prisoners with surnames beginning with “A”, “B”, “M” and “V” were included to increase the size of the sample. No differences were suspected between people whose names began with the arbitrarily chosen letters and other letters, and it was assumed that the distribution of time in prison was the same for these detainees as for the whole prison population. The day of entering the SIZO and the day of release was recorded for each person. No person-identifying information was kept.

We calculated the average temperature for each month in Saint Petersburg for the period 1996-2005 and chose the six coldest months (17). Analysis of seasonal variations in incidence was made for the date of entry to remand prisons and for date of TB development. We compared incidence between cases of developed disease for prisoners who entered remand prison during the cold period (CP) defined as the months from 1st November to 30th April, during the warm period (WP) from 1st May to 31st October, and also independent of entry time.

Information on the 466 prisoners who developed tuberculosis in SIZO-1 during the period 2000-2004 and in SIZO-4 during the period 2000-2002 was collected from the prison medical records. The average number of prisoners (per quarter during the year) in the SIZOs was also collected for each year. For each case the date of entry, date TB was first suspected based on chest x-ray and age was recorded. No other person-identifying information was kept. The diagnosis of TB was not made by x-ray only, it usually was confirmed by other diagnostic procedures in several weeks or months later. However, in order to get as close to the first signs of disease as possible, the date of x-ray was used as the date of falling ill.

Thereafter, all persons among the 360 TB cases who had developed the disease during the first year of imprisonment for the period 2000-2002 at both SIZOs were included. From this sample patients who had stayed less than one year in prison were chosen and divided into two groups: entering during the cold period and during the warm period (Table 1 and Table 2).


The TB incidence was defined as number of cases during the exact period, divided by the sum of person-months for cases and the sample, renormalized for the whole prison population. The proportion was multiplied by 12 to convert person-months to person-year. The incidence was calculated per 100 000.

For the calculation of incidence independent of time of entering prison, the average number prisoners in the SIZO (by quarter during the year) and the number of cases first diagnosed with chest x-ray was used. Comparison of the two incidence rates was made for each year and for each SIZO.

The proportion of prisoners who did not develop TB was calculated as a function of time since detention using Kaplan- Maier survival analysis.

The Nelson-Aalen estimator, a non-parametric method, was used to calculate the probability of getting TB for prisoners who had developed the disease during a certain 3-month period of time, given the number of prisoners who were healthy at the beginning of this time period.


The time served in remand prison varied from 1 to 88 months (mean: 9.8 months) during 2000-2002 (Table 1). The duration of stay in remand prison before TB development varied from 1 to 72 months (mean: 14.9 months) for the same 3-year period (Table 2).

The average length of stay for the prisoners who had spent less than one year in the SIZOs was 5.3 months. There was no difference in duration of imprisonment between people who were incarcerated during the CP or the WP (means: 5.2 and 5.4, respectively; p >0.05). The difference in means among TB patients who had entered prison in the CP versus the WP was not statistically significant: 14.5 and 15.6 months, respectively (p >0.05). Medians were 12 and 13 months. No difference was found in the duration of time between date of entry and date of developing TB for prisoners who had entered the prison during the CP or the WP and developed the disease during the first year. The period was 7.4 and 7.1 months, respectively.

The incidence of TB among prisoners who had stayed in remand prison less than 1 year and had entered prison in the CP was higher than the TB incidence for prisoners who had entered in the WP during 2000-2002 for both SIZOs: 1 953/100 000 person-year (95% confidence interval (CI) 1847; 2059) and 1123/100000 person-year (95% CI 1 041; 1 205), Risk ratio (RR) = 1.7 (95% CI 1.3; 2.4), respectively.

Analysis showed that fewer cases were registered during the period from April to September. However, no significant differences in TB incidence were observed between October to March (fourth and first quarters) when compared with April to September (second and third quarters) during the 5 years 2000-2004 (p>0.05).
 The survival analysis (“Kaplan-Maier”) is presented in Figure 1.

During the first 6 months the risk of getting TB was very small and changed only slightly with time. After the first 6 months there was a significant increase in the risk and this increase persisted with nearly the same rate after 12 months. More specifically, the integrated hazards for different time periods were calculated. Figure 2 shows the probability of getting TB at a certain time period among all prisoners who had developed TB given the number of persons who were healthy at the beginning of this time period (time interval - 3 months). Months 4-6 were used as the reference period and 95% CI were calculated.

The risk of getting TB increased during the first 2 years of imprisonment. Later the risk was probably constant at a high level (Figure 2).

The decline in risk during the period of 27 – 31 months after entry to the SIZOs may be explained by the fact that there were already very few cases registered.


The date of chest x-ray was used as the first sign of disease. A chest x-ray of every prisoner was taken every 6 months. All cases of suspected TB were confirmed clinically and the persons received anti-tuberculosis test therapy that could last up to two months.

Using the random samples collected for each SIZO for each year and TB cases (Table 1 and Table 2), it is not possible to compare duration in remand prison among all prisoners and among those who developed TB. For TB patients, the duration before disease was registered and for prisoners in general the entire time of imprisonment was registered. Regardless, patients who developed TB during imprisonment had been in the SIZOs longer (more than one year) than prisoners without TB. Consequently, limiting the time spent in remand prison to 6 months has an impact on health with respect to tuberculosis. Additional studies investigating how other diseases (including suicide) are related to duration of imprisonment are warranted.

The number of prisoners and the number of inmates in one cell increases slightly in winter. It is known that some homeless persons (or other people) intentionally commit crimes in order to be in prison during the cold winter. To maintain a constant room temperature the prisoners usually close all holes in the windows making ventilation impossible during day and night. This situation is advantageous for the spread of infectious diseases including tuberculosis. However, the TB incidence during the “winter” (October - March) was the same as during the “summer” (April - September) independent of the duration of stay in the SIZOs.

The risk of getting TB will increase exponentially if there is a spread of infection in the SIZOs. In our study, due to a relatively short period of stay and perhaps some other reason such as activation of latent infection, the risk is a logarithmic function (Figure 2). During the first 6 months, a relatively small number of prisoners developed the disease in remand prison. These few cases diagnosed may be a result of incorrect diagnosis at entry and quick development of disease during the 2-3 months after entering the prison. The increased number of prisoners who developed the disease after 6 months may be explained by reactivation of latent infection, relapse or becoming infected by other prisoners who had not been diagnosed.

It has been estimated that one year of jail time in the New York City jail system significantly increased the odds ratio of TB development to 2.2 (95% confidence interval, 1.1 to 4.4) (18). This calculation was made for convicted incarcerated persons. For prisoners in remand prison for more than 2 years it is not obvious whether the risk of getting TB increased with length of imprisonment. The maximum time spent in prison among patients who had developed TB was 72 months, but the number of prisoners who had spent 5-6 years in remand prison was small. Therefore it was not possible to calculate the risk dependent on time spent in prison for prisoners who had been in SIZOs more than 2 years due to the relatively small number of TB cases and prisoners in general who had spent so much time in prison.
We calculated that, approximately one of every 150 persons in remand prison (SIZO-1) in Saint Petersburg developed TB during imprisonment (2004). If prisoners are held in remand prison according to the new order, the number of new TB cases will be minimized. This will in turn reduce the impact of asymptomatic TB infection and transmission among prisoners. Prisoners who leave the SIZOs may contract TB later in a detention colony. Hence the total number of people who develop TB while in the prison system may remain the same. However, since conditions in detention colonies are essentially better than in remand, a decrease in the total number of people who developed TB is expected.

 According to Russian legislation, prisoners on remand should be held in a SIZO for no more than 12 months. However, in reality some prisoners were in remand prison for more than two years due to delays in the Court of Justice. In 2002 the legislation was revised reducing the maximum time period in remand to 6 months. Therefore the TB incidence among the prisoners who had stayed in the SIZO for less than 12 months was calculated. In this study the incidence of TB was significantly higher for prisoners who entered during the cold period. Therefore, more careful investigation of TB transmission and development of disease must be conducted among prisoners who enter prison during the period from November to April. 


1. Yurieva M, Sazhin V. TB prevalence in prison system. Zdravoohranenie Severo-Zapada 2002; 1: 61-64 (in Russian)
2. Annual reports from Institute of Phthisiopulmonology, St.Petersburg, Russia. 2003, 2004 “ (in Russian)
3. Global tuberculosis control-surveillance, planning, financing. WHO report 2005
4. Russo E. Turning Back the Tuberculosis Tide. The Scientist 23 May 2005, 19(10): 22
5. Bobrik A, Danishevski K, Eroshina K et al, Prison health in Russia: The Larger Picture. Journal of Public Health Policy 2005; 26: 30-59.
6. Punga V, Kapkov L. Tuberculosis in Russia. Probl Tuberk 1999; 1: 14-16 (in Russian)
7. Lobacheva T, Sazhin V, Vdovichenko E et al. Pulmonary tuberculosis in two remand prisons (SIZOs) in St.Petersburg, Russia. Eurosurveillance 2005; 10, 6: 93-96
8. Annual report for internal use. Department of the Penal System of Ministry of Justice of the Russian Federation in St. Petersburg and Leningrad region, St. Petersburg, Russia, 2005 (in Russian)
9. Koretskaya N. Risk factors for tuberculosis, specific features of its detection and course. Probl Tuberk 2002; 8:7-10 (in Russian)
10. Creation of the risk groups for pulmonary and extra-pulmonary tuberculosis. Recommendation N 2000/25. Probl Tuberk 2002; 10: 47-55 (in Russian)
11. Balabanova Ya M, Raddi N, Gram K et al. Analysis of risk factors of the occurrence of drug resistance in patients with tuberculosis from civil and penitentiary sectors in the Samara region. Probl Tuberk 2005; 5: 25-31 (in Russian)
12. Thorpe L, Frieden T, Laserson K et al. Seasonality of tuberculosis in India: is it real and what does it tell us? The Lancet 2004; 364: 1613-4
13. Rios M, Garcia J, Sanchez J et al. A statistical analysis of the seasonality in pulmonary tuberculosis. Eur J of Epidemiol 2000; 16: 483-488
14. Douglas A, Ali S, Bakhshi S. Does vitamin D deficiency account for ethic differences in tuberculosis seasonality in the UK? Ethn Health 1998; 3(4): 247-53
15. Douglas A, Strachan D, Maxwell J. Seasonality of tuberculosis: the reverse of other respiratory diseases in UK. Thorax 1996; 51(9): 944-6
16. Shilova M, Glumnaja T. Influence of seasonal and ecological factors on tuberculosis incidence. Probl Tuberk 2004; 2: 17-21 (in Russian)
17. Official Statistics and Weather Underground (
18. Bellin E, Fletcher D, Safyer S. Association of tuberculosis infection with increased time in or admission to the New York City jail system. JAMA 1993 May 5; 269(17): 2228-31